官网链接：http://iccv2021.thecvf.com/home
开会时间：2021年10月11日至17日

:exclamation::exclamation::exclamation::star2::star2::star2:📗📗📗ICCV 2021收录论文已全部公布，下载可在【我爱计算机视觉】后台回复“paper”，即可收到。共计 1612 篇。

:exclamation::exclamation::exclamation::star2::star2::star2:全部论文已粗略分类完毕，请查阅

历年综述论文分类汇总戳这里↘️CV-Surveys施工中~~~~~~~~~~

2022 年论文分类汇总戳这里

↘️CVPR-2022-Papers ↘️WACV-2022-Papers

2021年论文分类汇总戳这里

↘️ICCV-2021-Papers ↘️CVPR-2021-Papers

2020 年论文分类汇总戳这里

↘️CVPR-2020-Papers ↘️ECCV-2020-Papers

目录

:dog:	:mouse:	:hamster:	:tiger:
65.Optical Flow Estimation(光流估计)
61.Metric Learning(元学习)	62.Open-Set Recognition(开放集识别)	63.Data Augmentation(数据增强)	64.Anomaly Detection(异常检测)
57.Image Matching(图像匹配)	58.Computational Photography(光学、几何、光场成像、计算摄影)	59.Graph Neural Networks(图神经网络)	60.Federated Learning(联合学习)
53.Vision Localization(视觉定位)	54.Sketch recognition(草图)	55.Activity Recognition(活动识别)	56.Dataset(数据集)
49.Human-Object Interaction(人物交互)	50.Continual Learning(持续学习)	51.View Synthesis(视图合成)	52.Vision-and-Language(视觉语言)
45.Image Caption(图像字幕)	46.Defect Detection(缺陷检测)	47.NAS	48.6DoF
41.Out-of-Distribution Detection(OOD)	42.Visual Representations Learning(视觉表征学习)	43.Dense Prediction(密集预测)	44.Human motion prediction(人体运动预测)
37.Multitask Learning(多任务学习)	38.Weakly/Semi-Supervised/Self-supervised/Unsupervised Learning(自/半/弱监督学习)	39.Incremental Learning(增量学习)	40.Metric Learning(度量学习)
33.Remote Sensing Images(遥感影像)	34.Image Super-Resolution(图像超分辨率)	35.Quantization/Pruning/Knowledge Distillation/Model Compression(量化、剪枝、蒸馏、模型压缩/扩展与优化)	36.SLAM/AR/VR/机器人
29.Image Retrieval(图像检索)	30.Image Generation/synthesis(图像生成/合成)	31.Style Transfer(风格迁移)	32.语音
25.Medical Image(医学影像)	26.Image Processing(图像处理)	27.Multi-label image recognition(多标签图像识别)	28.Contrastive Learning(对比学习)]
21.Active Learning(主动学习)	22.GAN	23.Gaze Estimation(视线估计)	24.Face(人脸)
17.3D(三维视觉)	18.Transformers	19.Self-Driving Vehicles(自动驾驶)	20.Adversarial Learning(对抗学习)
13.Image Segmentation(图像分割)	14.Object Detection(目标检测)	15.Object Tracking(目标跟踪)	16.Re-Identification(重识别)
9.Video	10.OCR	11.Visual Question Answering(视觉问答)	12.Image/Fine-Grained Classification(图像/细粒度分类)
5.Few-Shot/Zero-Shot Learning;Domain Generalization/Adaptation(小/零样本学习;域适应/泛化)	6.Point Cloud(点云)	7.Scene Graph Generation(场景图生成)	8.Human Pose Estimation(人体姿态估计)
1.Other(其它)	2.Sign Language(手语识别)	3.Image Clustering(图像聚类)	4.Neural rendering(神经渲染)

65.Optical Flow Estimation(光流估计)

• Separable Flow: Learning Motion Cost Volumes for Optical Flow Estimation
  :star:code
• High-Resolution Optical Flow from 1D Attention and Correlation
  :open_mouth:oral:star:code
• GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning
  :star:code
• Sensor-Guided Optical Flow
  :star:code

64.Anomaly Detection(异常检测)

• 表面异常检测
  • DRÆM – A discriminatively trained reconstru
    :star:code
• 异常检测
  • Weakly Supervised Temporal Anomaly Segmentation with Dynamic Time Warping
  • Learning Unsupervised Metaformer for Anomaly Detection
    解决图像异常的分类或定位

63.Data Augmentation(数据增强)

• DivAug: Plug-In Automated Data Augmentation With Explicit Diversity Maximization
  :star:code
• TrivialAugment: Tuning-Free Yet State-of-the-Art Data Augmentation
  :open_mouth:oral:star:code
• Semantic Aware Data Augmentation for Cell Nuclei Microscopical Images With Artificial Neural Networks
• A Simple Baseline for Semi-Supervised Semantic Segmentation With Strong Data Augmentation

62.Open-Set Recognition(开放集识别)

• OpenGAN: Open-Set Recognition via Open Data Generation
  :trophy:Best Paper Honorable Mention
• Conditional Variational Capsule Network for Open Set Recognition
  :star:code

61.Metric Learning(元学习)

• Do Different Deep Metric Learning Losses Lead to Similar Learned Features?
  :star:code
• Learning With Memory-Based Virtual Classes for Deep Metric Learning
  :star:code

60.Federated Learning(联合学习)

• Federated Learning for Non-IID Data via Unified Feature Learning and Optimization Objective Alignment
• Ensemble Attention Distillation for Privacy-Preserving Federated Learning

59.Graph Neural Networks(图神经网络)

• Meta-Aggregator: Learning to Aggregate for 1-bit Graph Neural Networks
• PoGO-Net: Pose Graph Optimization With Graph Neural Networks
  :star:code
• Dynamic Dual Gating Neural Networks
  :star:code

58.Computational Photography(光学、几何、光场成像、计算摄影)

• An Asynchronous Kalman Filter for Hybrid Event Cameras
  :star:code
• 4D Cloud Scattering Tomography
• Snapshot compressive imaging(快照压缩成像)
  • Dense Deep Unfolding Network with 3D-CNN Prior for Snapshot Compressive Imaging
    :star:code
• 光场
  • Light Field Saliency Detection with Dual Local Graph Learning andReciprocative Guidance
  • Fast Light-Field Disparity Estimation With Multi-Disparity-Scale Cost Aggregation
    :star:code
  • SeLFVi: Self-supervised Light-Field Video Reconstruction from Stereo Video
  • SIGNET: Efficient Neural Representation for Light Fields
  • 光场重建
    • Learning Dynamic Interpolation for Extremely Sparse Light Fields With Wide Baselines
      :star:code
• 压缩成像
  • Time-Multiplexed Coded Aperture Imaging: Learned Coded Aperture and Pixel Exposures for Compressive Imaging Systems
• Homography Estimation
  • LocalTrans: A Multiscale Local Transformer Network for Cross-Resolution Homography Estimation
• 计算成像
  • Extreme-Quality Computational Imaging via Degradation Framework
• 光学像差矫正
  • Universal and Flexible Optical Aberration Correction Using Deep-Prior Based Deconvolution
    :star:code

57.Image Matching(图像匹配)

• Matching in the Dark: A Dataset for Matching Image Pairs of Low-light Scenes
• 特征点匹配
  • P2-Net: Joint Description and Detection of Local Features for Pixel and Point Matching
    :star:code

56.Dataset(数据集)

• Large Scale Multi-Illuminant (LSMI) Dataset for Developing White Balance Algorithm Under Mixed Illumination
  :sunflower:dataset
• FloW: A Dataset and Benchmark for Floating Waste Detection in Inland Waters
  :sunflower:dataset
  内陆水域漂浮废物检测数据集和基准
• FloorPlanCAD: A Large-Scale CAD Drawing Dataset for Panoptic Symbol Spotting
  :house:project
• 生物医学图像
  • BioFors: A Large Biomedical Image Forensics Dataset
• 3D重建
  • Common Objects in 3D: Large-Scale Learning and Evaluation of Real-life 3D Category Reconstruction
    :sunflower:dataset
• 航空影像数据集
  • Beyond Road Extraction: A Dataset for Map Update using Aerial Images
    :star:code:house:project
    用于使用航拍图像更新地图的数据集
• 动作识别
  • HAA500: Human-Centric Atomic Action Dataset with Curated Videos
• 目标识别
  • ORBIT: A Real-World Few-Shot Dataset for Teachable Object Recognition
    :star:code:sunflower:dataset
• 车道线检测
  • VIL-100: A New Dataset and a Baseline Model for Video Instance Lane Detection
    :sunflower:dataset
• 自动驾驶
  • Large Scale Interactive Motion Forecasting for Autonomous Driving: The Waymo Open Motion Dataset
• 视觉语言数据集
  • E-ViL: A Dataset and Benchmark for Natural Language Explanations in Vision-Language Tasks
    :star:codeVL
• DeepFake检测
  • KoDF: A Large-Scale Korean DeepFake Detection Dataset
    :sunflower:dataset
• 高质量视频
  • Seeing Dynamic Scene in the Dark: A High-Quality Video Dataset With Mechatronic Alignment
    :sunflower:dataset视频

55.Activity Recognition(活动识别)

• Selective Feature Compression for Efficient Activity Recognition Inference
• 小组活动识别
  • Spatio-Temporal Dynamic Inference Network for Group Activity Recognition
    :star:code
  • GroupFormer: Group Activity Recognition with Clustered Spatial-Temporal Transformer
    :star:code

54.Sketch recognition(草图)

• SketchLattice: Latticed Representation for Sketch Manipulation
• SketchAA: Abstract Representation for Abstract Sketches

53.Vision Localization(视觉定位)

• Continual Learning for Image-Based Camera Localization
  :star:code
• CrowdDriven: A New Challenging Dataset for Outdoor Visual Localization
  :sunflower:dataset
• Pose Correction for Highly Accurate Visual Localization in Large-Scale Indoor Spaces
  :star:code
• Cross-Descriptor Visual Localization and Mapping

52.Vision-and-Language(视觉语言)

• YouRefIt: Embodied Reference Understanding with Language and Gesture
  :open_mouth:oral:house:project
• VLGrammar: Grounded Grammar Induction of Vision and Language
  :star:code
• COOKIE: Contrastive Cross-Modal Knowledge Sharing Pre-Training for Vision-Language Representation
  :star:code
• Panoptic Narrative Grounding
  :open_mouth:oral:star:code
• AESOP: Abstract Encoding of Stories, Objects, and Pictures
  :star:code:tv:video
• Adaptive Hierarchical Graph Reasoning With Semantic Coherence for Video-and-Language Inference
• 视觉推理
  • Interpretable Visual Reasoning via Induced Symbolic Space
• 语义导航
  • THDA: Treasure Hunt Data Augmentation for Semantic Navigation
• 视觉语言导航
  • Airbert: In-domain Pretraining for Vision-and-Language Navigation
    :house:project
  • Waypoint Models for Instruction-guided Navigation in Continuous Environments
    :open_mouth:oral:star:code:house:project:tv:video
  • The Road To Know-Where: An Object-and-Room Informed Sequential BERT for Indoor Vision-Language Navigation
    :star:code
  • Vision-Language Navigation With Random Environmental Mixup
• 视觉对话导航
  • Self-Motivated Communication Agent for Real-World Vision-Dialog Navigation
• 视觉导航
  • Pose Invariant Topological Memory for Visual Navigation
    :star:code
  • Visual Graph Memory With Unsupervised Representation for Visual Navigation
    :star:code:house:project:tv:video
• visual grounding
  • InstanceRefer: Cooperative Holistic Understanding for Visual Grounding on Point Clouds through Instance Multi-level Contextual Referring
    :star:code
  • TransVG: End-to-End Visual Grounding With Transformers
    :star:code
• 视觉对话
  • Unified Questioner Transformer for Descriptive Question Generation in Goal-Oriented Visual Dialogue

51.View Synthesis(视图合成)

• Out-of-boundary View Synthesis Towards Full-Frame Video Stabilization
  :star:code
• Deep 3D Mask Volume for View Synthesis of Dynamic Scenes
  :house:project
• Embedding Novel Views in a Single JPEG Image
• Video Autoencoder: self-supervised disentanglement of static 3D structure and motion
  :open_mouth:oral:star:code:house:project:tv:video
• Geometry-Free View Synthesis: Transformers and No 3D Priors
  :star:code
• Dynamic View Synthesis From Dynamic Monocular Video
  :house:project:tv:video
• Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis
  :house:project:tv:video
• Infinite Nature: Perpetual View Generation of Natural Scenes from a Single Image
  :open_mouth:oral:star:code:house:project:tv:video
• Worldsheet: Wrapping the World in a 3D Sheet for View Synthesis from a Single Image
  :open_mouth:oral:star:code:house:project:tv:video

50.Continual Learning(持续学习)

• Online Continual Learning with Natural Distribution Shifts: An Empirical Study with Visual Data
  :star:code
• Continual Learning on Noisy Data Streams via Self-Purified Replay
  :star:code
• Rehearsal Revealed: The Limits and Merits of Revisiting Samples in Continual Learning
  :star:code
• Co2L: Contrastive Continual Learning
  :star:code

49.Human-Object Interaction(人物交互)

• Exploiting Scene Graphs for Human-Object Interaction Detection
  :star:code
• Spatially Conditioned Graphs for Detecting Human-Object Interactions
  :star:code:tv:video
• Virtual Multi-Modality Self-Supervised Foreground Matting for Human-Object Interaction
• Detecting Human-Object Relationships in Videos
• Weakly Supervised Human-Object Interaction Detection in Video via Contrastive Spatiotemporal Regions
  :star:code:house:project:sunflower:dataset
• Discovering Human Interactions With Large-Vocabulary Objects via Query and Multi-Scale Detection
  :star:code
• Visual Relationship Detection Using Part-and-Sum Transformers With Composite QueriesVRD和HOI
• Interaction Compass: Multi-Label Zero-Shot Learning of Human-Object Interactions via Spatial Relations
  :star:code
• H2O
  • H2O: A Benchmark for Visual Human-Human Object Handover Analysis
• Human Interaction Understanding
  • Consistency-Aware Graph Network for Human Interaction Understanding
    :star:code
  • H2O: Two Hands Manipulating Objects for First Person Interaction Recognition
    :house:project
• 手物交互
  • Toward Human-Like Grasp: Dexterous Grasping via Semantic Representation of Object-Hand
  • Reconstructing Hand-Object Interactions in the Wild
    :house:project
  • CPF: Learning a Contact Potential Field To Model the Hand-Object Interaction
    :star:code手物交互
• HOI(行为理解)
  • GeomNet: A Neural Network Based on Riemannian Geometries of SPD Matrix Space and Cholesky Space for 3D Skeleton-Based Interaction Recognition

48.6DoF

• SO-Pose: Exploiting Self-Occlusion for Direct 6D Pose Estimation
  :star:code
• StereOBJ-1M: Large-scale Stereo Image Dataset for 6D Object Pose Estimation
• SGPA: Structure-Guided Prior Adaptation for Category-Level 6D Object Pose Estimation
• RePOSE: Fast 6D Object Pose Refinement via Deep Texture Rendering
  :star:code
• DualPoseNet: Category-Level 6D Object Pose and Size Estimation Using Dual Pose Network With Refined Learning of Pose Consistency
  :star:code
• PR-GCN: A Deep Graph Convolutional Network With Point Refinement for 6D Pose Estimation
• 物体姿势估计
  • CAPTRA: CAtegory-Level Pose Tracking for Rigid and Articulated Objects From Point Clouds
    :open_mouth:oral:star:code:house:project:tv:video

47.NAS

• BN-NAS: Neural Architecture Search with Batch Normalization
  :star:code
• RANK-NOSH: Efficient Predictor-Based Architecture Search via Non-Uniform Successive Halving
• Pi-NAS: Improving Neural Architecture Search by Reducing Supernet Training Consistency Shift
  :star:code
• Evolving Search Space for Neural Architecture Search
  :star:code:tv:video
• FairNAS: Rethinking Evaluation Fairness of Weight Sharing Neural Architecture Search
  :star:code
• GLiT: Neural Architecture Search for Global and Local Image Transformer
  :star:code
• Neural Architecture Search for Joint Human Parsing and Pose Estimation
  :star:code
• Distilling Optimal Neural Networks: Rapid Search in Diverse Spaces
• Learning Latent Architectural Distribution in Differentiable Neural Architecture Search via Variational Information Maximization
• Not All Operations Contribute Equally: Hierarchical Operation-Adaptive Predictor for Neural Architecture Search
• Zen-NAS: A Zero-Shot NAS for High-Performance Image Recognition
  :star:code
• BossNAS: Exploring Hybrid CNN-Transformers With Block-Wisely Self-Supervised Neural Architecture Search
  :star:code
• NAS-OoD: Neural Architecture Search for Out-of-Distribution Generalization
• AutoSpace: Neural Architecture Search With Less Human Interference
  :star:code
• IDARTS: Interactive Differentiable Architecture Search

46.Defect Detection(缺陷检测)

• DRÆM -- A discriminatively trained reconstruction embedding for surface anomaly detection

45.Image Caption(图像字幕)

• Who's Waldo? Linking People Across Text and Images
  :open_mouth:oral:house:project
  :newspaper:解读:ICCV2021 Oral-新任务！新数据集！康奈尔大学提出了类似VG但又不是VG的PVG任务
• Partial Off-Policy Learning: Balance Accuracy and Diversity for Human-Oriented Image Captioning
• Topic Scene Graph Generation by Attention Distillation From Caption
  :star:code
• Understanding and Evaluating Racial Biases in Image Captioning
  :star:code:house:project
• In Defense of Scene Graphs for Image Captioning
  :star:code
• art description generation(艺术描述生成)
  • Explain Me the Painting: Multi-Topic Knowledgeable Art Description Generation
    :star:code
• Change Captioning
  • Viewpoint-Agnostic Change Captioning With Cycle Consistency

44.Human motion prediction(人体运动预测)

• MSR-GCN: Multi-Scale Residual Graph Convolution Networks for Human Motion Prediction
  :star:code
• Stochastic Scene-Aware Motion Prediction
  :star:code:house:project
• Generating Smooth Pose Sequences for Diverse Human Motion Prediction
  :open_mouth:oral:star:code
• TRiPOD: Human Trajectory and Pose Dynamics Forecasting in the Wild
  :house:project
• Motion Prediction using Trajectory Cues
• 3D人体运动预测
  • Contextually Plausible and Diverse 3D Human Motion Prediction

43.Dense Prediction(密集预测)

• FaPN: Feature-aligned Pyramid Network for Dense Image Prediction
  :star:code
• 多任务密集预测
  • Exploring Relational Context for Multi-Task Dense Prediction

42.Representations Learning(表征学习)

• Learning From Noisy Data With Robust Representation Learning
  :star:code
• Self-Supervised Representation Learning From Flow Equivariance
• Exploring Visual Engagement Signals for Representation Learning
  :star:code
• Switchable K-class Hyperplanes for Noise-Robust Representation Learning
  :star:code
• Region Similarity Representation Learning
  :star:code
• Curious Representation Learning for Embodied Intelligence
  :star:code:house:project
• 视觉表征学习
  • Self-Supervised Visual Representations Learning by Contrastive Mask Prediction
    :newspaper:解读:ICCV2021 比MoCo更通用的对比学习范式，中科大&MSRA提出对比学习新方法MaskCo
  • Temporal Knowledge Consistency for Unsupervised Visual Representation Learning
  • Contrasting Contrastive Self-Supervised Representation Learning Pipelines
    :star:code
  • Concept Generalization in Visual Representation Learning
    :house:project
  • Collaborative Unsupervised Visual Representation Learning from Decentralized Data
  • Episodic Transformer for Vision-and-Language Navigation
    :star:code
  • Multi-VAE: Learning Disentangled View-Common and View-Peculiar Visual Representations for Multi-View Clustering
• 视频表示学习
  • Composable Augmentation Encoding for Video Representation Learning
  • Motion-Focused Contrastive Learning of Video Representations
• ASCNet: Self-Supervised Video Representation Learning With Appearance-Speed Consistency
• ACAV100M: Automatic Curation of Large-Scale Datasets for Audio-Visual Video Representation Learning
  :house:project
• Time-Equivariant Contrastive Video Representation Learning
• Space-Time Crop & Attend: Improving Cross-Modal Video Representation Learning
  :star:code

41.Out-of-Distribution Detection(OOD)

• CODEs: Chamfer Out-of-Distribution Examples against Overconfidence Issue
• Semantically Coherent Out-of-Distribution Detection
  :star:code:house:project
• The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization
  :star:code

40.Metric Learning(度量学习)

• Towards Interpretable Deep Metric Learning with Structural Matching
  :star:code
• Deep Relational Metric Learning
  :star:code
• LoOp: Looking for Optimal Hard Negative Embeddings for Deep Metric Learning
  :star:code
• Manifold Matching via Deep Metric Learning for Generative Modeling
  :star:code

39.Incremental Learning(增量学习)

• 类增量学习
  • Always Be Dreaming: A New Approach for Data-Free Class-Incremental Learning
    :newspaper:解读:让模型实现“终生学习”，佐治亚理工学院提出Data-Free的增量学习
  • Striking a Balance Between Stability and Plasticity for Class-Incremental Learning
  • Synthesized Feature Based Few-Shot Class-Incremental Learning on a Mixture of Subspaces
    :star:code

38.Weakly/Semi-Supervised/Self-supervised/Unsupervised Learning(自/半/弱监督学习)

• 半监督
  • Trash to Treasure: Harvesting OOD Data with Cross-Modal Matching for Open-Set Semi-Supervised Learning
  • Semi-Supervised Learning of Visual Features by Non-Parametrically Predicting View Assignments With Support Samples
    :star:code
  • Semi-Supervised Active Learning for Semi-Supervised Models: Exploit Adversarial Examples With Graph-Based Virtual Labels
  • CoMatch: Semi-Supervised Learning With Contrastive Graph Regularization
    :star:code
  • Multiview Pseudo-Labeling for Semi-supervised Learning from Video
• 自监督
  • Focus on the Positives: Self-Supervised Learning for Biodiversity Monitoring
    :star:code
  • Self-supervised Neural Networks for Spectral Snapshot Compressive Imaging
    :star:code
  • ISD: Self-Supervised Learning by Iterative Similarity Distillation
    :star:code
  • Contrast and Order Representations for Video Self-Supervised Learning
  • On Feature Decorrelation in Self-Supervised Learning
    :open_mouth:oral
  • Geography-Aware Self-Supervised Learning
  • Multimodal Clustering Networks for Self-supervised Learning from Unlabeled Videos
  • Efficient Visual Pretraining with Contrastive Detection
  • Broaden Your Views for Self-Supervised Video Learning
  • CDS: Cross-Domain Self-supervised Pre-training
  • On Compositions of Transformations in Contrastive Self-Supervised Learning
    :star:code
  • Solving Inefficiency of Self-Supervised Representation Learning
    :star:code
  • Divide and Contrast: Self-supervised Learning from Uncurated Data
  • Emerging Properties in Self-Supervised Vision Transformers
    :star:code
  • Mean Shift for Self-Supervised Learning
    :star:code
• 弱监督
  • Weakly Supervised Representation Learning With Coarse Labels
    :star:code

37.Multitask Learning(多任务学习)

• MultiTask-CenterNet (MCN): Efficient and Diverse Multitask Learning using an Anchor Free Approach
  :newspaper:解读:ICCV2021《MultiTask CenterNet》CV多任务新进展！一节更比三节强
• Multi-Task Self-Training for Learning General Representations
  :newspaper:解读:ICCV2021 MuST：还在特定任务里为刷点而苦苦挣扎？谷歌的大佬们都已经开始玩多任务训练了
• UniT: Multimodal Multitask Learning With a Unified Transformer
  :star:code
• Learning Multiple Pixelwise Tasks Based on Loss Scale Balancing
  :star:code
• Learning With Privileged Tasks
• Task Switching Network for Multi-Task Learning

36.SLAM/AR/VR/机器人

• 机器人
  • 室内导航
    • The Surprising Effectiveness of Visual Odometry Techniques for Embodied PointGoal Navigation
      :star:code:house:project
    • Pathdreamer: A World Model for Indoor Navigation
      :tv:video
  • 机器手抓取
    • Hand-Object Contact Consistency Reasoning for Human Grasps Generation
      :open_mouth:oral:star:code:house:project:tv:video
• VR/AR
  • The Power of Points for Modeling Humans in Clothing
    :star:code:house:project:tv:video
  • 虚拟试穿
    • M3D-VTON: A Monocular-to-3D Virtual Try-On Network
      :star:code
    • ZFlow: Gated Appearance Flow-based Virtual Try-on with 3D Priors
    • Dressing in Order: Recurrent Person Image Generation for Pose Transfer, Virtual Try-On and Outfit Editing
    • FashionMirror: Co-Attention Feature-Remapping Virtual Try-On With Sequential Template Poses
    • Structure-transformed Texture-enhanced Network for Person Image Synthesis
• SLAM
  • On the Limits of Pseudo Ground Truth in Visual Camera Re-localisation
    :star:code
  • Transfusion: A Novel SLAM Method Focused on Transparent Objects
  • iMAP: Implicit Mapping and Positioning in Real-Time
  • Learning To Bundle-Adjust: A Graph Network Approach to Faster Optimization of Bundle Adjustment for Vehicular SLAM
  • R-SLAM: Optimizing Eye Tracking From Rolling Shutter Video of the Retina
  • Place Recognition
    • Attentional Pyramid Pooling of Salient Visual Residuals for Place Recognition
    • Pyramid Point Cloud Transformer for Large-Scale Place Recognition
      :star:code

35.Quantization/Pruning/Knowledge Distillation/Model Compression(量化、剪枝、蒸馏、模型压缩/扩展与优化)

• 知识蒸馏
  • Distilling Holistic Knowledge with Graph Neural Networks
    :star:code
  • Lipschitz Continuity Guided Knowledge Distillation
    :star:code
  • Densely Guided Knowledge Distillation Using Multiple Teacher Assistants
    :star:code
  • Revisiting Adversarial Robustness Distillation: Robust Soft Labels Make Student Better
    :star:code
  • Compressing Visual-linguistic Model via Knowledge Distillation
  • Self-Knowledge Distillation With Progressive Refinement of Targets
    :star:code:tv:video
  • Student Customized Knowledge Distillation: Bridging the Gap Between Student and Teacher
  • Channel-Wise Knowledge Distillation for Dense Prediction
    :star:code
  • Exploring Inter-Channel Correlation for Diversity-Preserved Knowledge Distillation
    :star:code
• 量化
  • Distance-aware Quantization
    :star:code:house:project
  • Dynamic Network Quantization for Efficient Video Inference
    :star:code:house:project
  • Cluster-Promoting Quantization with Bit-Drop for Minimizing Network Quantization Loss
  • Improving Low-Precision Network Quantization via Bin Regularization
  • Towards Mixed-Precision Quantization of Neural Networks via Constrained Optimization
  • Integer-arithmetic-only Certified Robustness for Quantized Neural Networks
  • RMSMP: A Novel Deep Neural Network Quantization Framework With Row-Wise Mixed Schemes and Multiple Precisions
  • Improving Neural Network Efficiency via Post-Training Quantization With Adaptive Floating-Point
    :star:code
  • Once Quantization-Aware Training: High Performance Extremely Low-bit Architecture Search
    :star:code
• 模型压缩
  • GDP: Stabilized Neural Network Pruning via Gates with Differentiable Polarization
  • Exploration and Estimation for Model Compression
• 剪枝
  • ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgetting
    :star:code
  • Auto Graph Encoder-Decoder for Neural Network Pruning

34.Super-Resolution(超分辨率)

• ISR
  • Mutual Affine Network for Spatially Variant Kernel Estimation in Blind Image Super-Resolution
    :star:code
  • Hierarchical Conditional Flow: A Unified Framework for Image Super-Resolution and Image Rescaling
    :star:code
  • Deep Reparametrization of Multi-Frame Super-Resolution and Denoising
    :open_mouth:oral
  • Dual-Camera Super-Resolution with Aligned Attention Modules
    :star:code:house:project:tv:video
  • Attention-Based Multi-Reference Learning for Image Super-Resolution
    :star:code:house:project
  • Learning a Single Network for Scale-Arbitrary Super-Resolution
  • Fourier Space Losses for Efficient Perceptual Image Super-Resolution
    :star:code
  • Achieving On-Mobile Real-Time Super-Resolution With Neural Architecture and Pruning Search
  • Designing a Practical Degradation Model for Deep Blind Image Super-Resolution
    :star:code
  • Event Stream Super-Resolution via Spatiotemporal Constraint Learning
  • Dynamic High-Pass Filtering and Multi-Spectral Attention for Image Super-Resolution
  • Super-Resolving Cross-Domain Face Miniatures by Peeking at One-Shot Exemplar
  • Context Reasoning Attention Network for Image Super-Resolution
  • EvIntSR-Net: Event Guided Multiple Latent Frames Reconstruction and Super-Resolution
  • Super Resolve Dynamic Scene from Continuous Spike Streams
  • Deep Blind Video Super-Resolution
  • Benchmarking Ultra-High-Definition Image Super-Resolution
  • Lucas-Kanade Reloaded: End-to-End Super-Resolution From Raw Image Bursts
  • Unsupervised Real-World Super-Resolution: A Domain Adaptation Perspective
  • Real-World Video Super-Resolution: A Benchmark Dataset and a Decomposition Based Learning Scheme
    :star:code
    :newspaper:解读:ICCV2021 香港理工、阿里达摩院提出RealVSR：视频超分任务中的新数据集与损失方案
• VSR
  • Omniscient Video Super-Resolution
    :star:code
  • COMISR: Compression-Informed Video Super-Resolution
    :star:code
    :newspaper:解读:谷歌提出COMISR算法：针对视频压缩的压缩感知超分辨率
  • Learning Frequency-Aware Dynamic Network for Efficient Super-Resolution
  • Efficient Video Compression via Content-Adaptive Super-Resolution
    :star:code

33.Remote Sensing Images(遥感影像)

• SUNet: Symmetric Undistortion Network for Rolling Shutter Correction
• Change is Everywhere: Single-Temporal Supervised Object Change Detection in Remote Sensing Imagery
  :star:code
  :newspaper:解读:ICCV2021｜武汉大学RSIDEA团队提出一种新颖的弱监督遥感变化检测算法STAR
• 卫星图像全景视频合成
  • Sat2Vid: Street-view Panoramic Video Synthesis from a Single Satellite Image
• 基于卫星影像的交通事故检测
  • Inferring High-Resolution Traffic Accident Risk Maps Based on Satellite Imagery and GPS Trajectories
• 遥感数据
  • Seasonal Contrast: Unsupervised Pre-Training From Uncurated Remote Sensing Data
    :star:code
  • Dynamic Cross Feature Fusion for Remote Sensing Pansharpening
• 分割
  • Self-Mutating Network for Domain Adaptive Segmentation in Aerial Images
  • 卫星图像的全景分割
    • Panoptic Segmentation of Satellite Image Time Series With Convolutional Temporal Attention Networks
      :star:code:sunflower:PASTIS dataset
• 三维重建
  • 3D Building Reconstruction from Monocular Remote Sensing Images
    :house:project

32.语音

• The Right to Talk: An Audio-Visual Transformer Approach
  :star:code
• Image2Reverb: Cross-Modal Reverb Impulse Response Synthesis
  :star:code:house:project
• 音频分离
  • Visual Scene Graphs for Audio Source Separation
• 音频-手势
  • Audio2Gestures: Generating Diverse Gestures from Speech Audio with Conditional Variational Autoencoders
    :house:project
• Active Speaker Detection(ASD主动式扬声器检测)
  • How To Design a Three-Stage Architecture for Audio-Visual Active Speaker Detection in the Wild
    :star:code
  • MAAS: Multi-Modal Assignation for Active Speaker Detection
• 从人脸视频中重新收集音频
  • Multi-Modality Associative Bridging Through Memory: Speech Sound Recollected From Face Video
• 视听源定位
  • Localize to Binauralize: Audio Spatialization From Visual Sound Source Localization
    :star:code:tv:video
• 视听源分离
  • Move2Hear: Active Audio-Visual Source Separation
    :star:code:house:project
• 视听平面图重建
  • Audio-Visual Floorplan Reconstruction
    :star:code:house:project:tv:video

31.Style Transfer(风格迁移)

• AdaAttN: Revisit Attention Mechanism in Arbitrary Neural Style Transfer
  :star:code
• Domain-Aware Universal Style Transfer
  :star:code
• Diverse Image Style Transfer via Invertible Cross-Space Mapping
• StyleFormer: Real-Time Arbitrary Style Transfer via Parametric Style Composition
• Manifold Alignment for Semantically Aligned Style Transfer
  :star:code

30.Image Generation/synthesis(图像生成/合成)

• ILVR: Conditioning Method for Denoising Diffusion Probabilistic Models
  :open_mouth:oral
• Image Synthesis via Semantic Composition
  :star:code:house:project
• Image Synthesis From Layout With Locality-Aware Mask Adaption
• 图像融合
  • DTMNet: A Discrete Tchebichef Moments-Based Deep Neural Network for Multi-Focus Image Fusion

29.Image Retrieval(图像检索)

• DOLG: Single-Stage Image Retrieval with Deep Orthogonal Fusion of Local and Global Features
  :star:code
• Image Retrieval on Real-life Images with Pre-trained Vision-and-Language Models
  :star:code:house:project
• Self-supervised Product Quantization for Deep Unsupervised Image Retrieval
  :star:code
• Instance-Level Image Retrieval Using Reranking Transformers
  :star:code
• Learning Attribute-Driven Disentangled Representations for Interactive Fashion Retrieval
  :star:code
• Telling the What While Pointing to the Where: Multimodal Queries for Image Retrieval
• Frozen in Time: A Joint Video and Image Encoder for End-to-End Retrieval
• Learning Deep Local Features With Multiple Dynamic Attentions for Large-Scale Image Retrieval
  :star:code
• Bayesian Triplet Loss: Uncertainty Quantification in Image Retrieval
• 跨域检索
  • Universal Cross-Domain Retrieval: Generalizing Across Classes and Domains
• Visual Geolocalization
  • Viewpoint Invariant Dense Matching for Visual Geolocalization
    :star:code
• 跨模态检索
  • Ask&Confirm: Active Detail Enriching for Cross-Modal Retrieval With Partial Query
    :star:code
  • Product1M: Towards Weakly Supervised Instance-Level Product Retrieval via Cross-modal Pretraining
    :star:code
  • Wasserstein Coupled Graph Learning for Cross-Modal Retrieval
  • Adversarial Attack on Deep Cross-Modal Hamming Retrieval
• 文本-视频检索
  • TEACHTEXT: CrossModal Generalized Distillation for Text-Video Retrieval
    :house:project
• 视频- 文本检索
  • HiT: Hierarchical Transformer With Momentum Contrast for Video-Text Retrieval
• image-based 3D shape retrieval
  • Single Image 3D Shape Retrieval via Cross-Modal Instance and Category Contrastive Learning
• 近邻搜索
  • Product Quantizer Aware Inverted Index for Scalable Nearest Neighbor Search

28.Contrastive Learning(对比学习)

• Improving Contrastive Learning by Visualizing Feature Transformation
  :open_mouth:oral:star:code
• TACo: Token-aware Cascade Contrastive Learning for Video-Text Alignment
  :newspaper:解读:ICCV2021-TOCo-微软&CMU提出Token感知的级联对比学习方法，在视频文本对齐任务上“吊打”其他SOTA方法
• A Broad Study on the Transferability of Visual Representations With Contrastive Learning
  :star:code
• Vi2CLR: Video and Image for Visual Contrastive Learning of Representation
• LatentCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable Directions
  :star:code
• CrossCLR: Cross-Modal Contrastive Learning for Multi-Modal Video Representations
• Social NCE: Contrastive Learning of Socially-Aware Motion Representations
  :star:code:tv:video
• With a Little Help From My Friends: Nearest-Neighbor Contrastive Learning of Visual Representations
• Contrastive Learning of Image Representations With Cross-Video Cycle-Consistency
  :house:project
• Weakly Supervised Contrastive Learning

27.Multi-label image recognition(多标签图像识别)

• Residual Attention: A Simple but Effective Method for Multi-Label Recognition
  :star:code
• Transformer-based Dual Relation Graph for Multi-label Image Recognition

26.Image Processing(图像处理)

• Aligning Latent and Image Spaces to Connect the Unconnectable
  :star:code:house:project
• 图像形状操纵
  • Image Shape Manipulation from a Single Augmented Training Sample
    :open_mouth:oral:star:code:house:project
• 边缘检测
  • RINDNet: Edge Detection for Discontinuity in Reflectance, Illumination, Normal and Depth
    :open_mouth:oral:star:code
  • Pixel Difference Networks for Efficient Edge Detection
    :star:code
• 图像识别
  • MicroNet: Improving Image Recognition with Extremely Low FLOPs
    :star:code
• 图像去模糊
  • Rethinking Coarse-to-Fine Approach in Single Image Deblurring
    :star:code
  • Single Image Defocus Deblurring Using Kernel-Sharing Parallel Atrous Convolutions
  • Defocus Map Estimation and Deblurring From a Single Dual-Pixel Image
  • Motion Deblurring with Real Events
  • Pyramid Architecture Search for Real-Time Image Deblurring
  • 运动去模糊
    • Perceptual Variousness Motion Deblurring With Light Global Context Refinement
• 视频去模糊
  • Bringing Events Into Video Deblurring With Non-Consecutively Blurry Frames
    :star:code
• Image quality assessment(图像质量评估IQA)
  • MUSIQ: Multi-scale Image Quality Transformer
    :star:code
• Image Harmonization
  • SSH: A Self-Supervised Framework for Image Harmonization
    :star:code
  • Learning Conditional Knowledge Distillation for Degraded-Reference Image Quality Assessment
    :star:code
• 去阴影
  • CANet: A Context-Aware Network for Shadow Removal
    :star:code
  • DC-ShadowNet: Single-Image Hard and Soft Shadow Removal Using Unsupervised Domain-Classifier Guided Network
• 去噪
  • Rethinking Deep Image Prior for Denoising
    :star:code
  • Rethinking Noise Synthesis and Modeling in Raw Denoising
    :star:code
  • C2N: Practical Generative Noise Modeling for Real-World Denoising
  • The Benefit of Distraction: Denoising Camera-Based Physiological Measurements Using Inverse Attention
    :star:code
  • Hyperspectral Image Denoising with Realistic Data
    :star:code
  • End-to-End Unsupervised Document Image Blind Denoising
  • Cross-Patch Graph Convolutional Network for Image Denoising
  • 视频去噪
    • Patch Craft: Video Denoising by Deep Modeling and Patch Matching
• 图像着色
  • Towards Vivid and Diverse Image Colorization with Generative Color Prior
  • Deep Edge-Aware Interactive Colorization Against Color-Bleeding Effects
    :open_mouth:oral:house:project
• 图像增强
  • Real-time Image Enhancer via Learnable Spatial-aware 3D Lookup Tables
  • Adaptive Unfolding Total Variation Network for Low-Light Image Enhancement
    :star:code
  • Representative Color Transform for Image Enhancement
  • STAR: A Structure-Aware Lightweight Transformer for Real-Time Image Enhancement
  • Deep Symmetric Network for Underexposed Image Enhancement With Recurrent Attentional Learning
    :star:code:house:project
  • StarEnhancer: Learning Real-Time and Style-Aware Image Enhancement
• 图像恢复
  • Spatially-Adaptive Image Restoration using Distortion-Guided Networks
    :star:code
  • Dynamic Attentive Graph Learning for Image Restoration
    :star:code
  • Self-Supervised Cryo-Electron Tomography Volumetric Image Restoration From Single Noisy Volume With Sparsity Constraint
    :star:code
  • Searching for Controllable Image Restoration Networks
    :star:code
• 图像压缩
  • Variable-Rate Deep Image Compression through Spatially-Adaptive Feature Transform
    :star:code
  • Neural Image Compression via Attentional Multi-Scale Back Projection and Frequency Decomposition
• 图像修复
  • Image Inpainting via Conditional Texture and Structure Dual Generation
    :star:code
  • CR-Fill: Generative Image Inpainting With Auxiliary Contextual Reconstruction
    :star:code
  • Parallel Multi-Resolution Fusion Network for Image Inpainting
  • Painting from Part
    :star:code
  • WaveFill: A Wavelet-Based Generation Network for Image Inpainting
  • Distillation-Guided Image Inpainting
  • Learning a Sketch Tensor Space for Image Inpainting of Man-made Scenes
    :star:code:house:project
• Image extrapolation
  • SemIE: Semantically-aware Image Extrapolation
    :house:project
• Reversible Image Conversion
  • IICNet: A Generic Framework for Reversible Image Conversion
    :star:code
• 伪影去除
  • Towards Flexible Blind JPEG Artifacts Removal
    :star:code
  • Learning Dual Priors for JPEG Compression Artifacts Removal
  • Let's See Clearly: Contaminant Artifact Removal for Moving Cameras
• De-rendering
  • De-rendering Stylized Texts
    :star:code:house:project
• 去除光晕
  • Light Source Guided Single-Image Flare Removal From Unpaired Data
• 全景图拼接
  • Minimal Solutions for Panoramic Stitching Given Gravity Prior
• Flare Removal
  • How to Train Neural Networks for Flare Removal
    :house:project:tv:video
• 图像裁剪
  • TransView: Inside, Outside, and Across the Cropping View Boundaries
  • Dissecting Image Crops
    :star:code
• 去反射
  • Location-Aware Single Image Reflection Removal
    :star:code
  • V-DESIRR: Very Fast Deep Embedded Single Image Reflection Removal
    :star:code
• 去雨
  • Improving De-Raining Generalization via Neural Reorganization
  • Unpaired Learning for Deep Image Deraining with Rain Direction Regularize
    :house:project
  • Structure-Preserving Deraining With Residue Channel Prior Guidance
    :star:code
• 图像失真去除
  • Unsupervised Non-Rigid Image Distortion Removal via Grid Deformation
    :star:code:tv:video
• 消除水下图像的折射失真
  • Learning To Remove Refractive Distortions From Underwater Images
• 图像补全
  • High-Fidelity Pluralistic Image Completion With Transformers
    :star:code:house:project
• Image Decomposition
  • Unsupervised Layered Image Decomposition into Object Prototypes
• 失真矫正
  • Towards Complete Scene and Regular Shape for Distortion Rectification by Curve-Aware Extrapolation
• HDR
  • Unpaired Learning for High Dynamic Range Image Tone Mapping
  • 超高清图像HDR重建
    • Ultra-High-Definition Image HDR Reconstruction via Collaborative Bilateral Learning
• 图像去雪
  • ALL Snow Removed: Single Image Desnowing Algorithm Using Hierarchical Dual-Tree Complex Wavelet Representation and Contradict Channel Loss
    :star:code
• Image Harmonization
  • Image Harmonization With Transformer
    :star:code
• 图像编辑
  • Language-Guided Global Image Editing via Cross-Modal Cyclic Mechanism
• image hiding(图像隐藏)
  • HiNet: Deep Image Hiding by Invertible Network
    :star:code

25.Medical Image(医学影像)

• Equivariant Imaging: Learning Beyond the Range Space
  :open_mouth:oral:star:code
• Deep Survival Analysis With Longitudinal X-Rays for COVID-19
• 医学图像分割
  • Recurrent Mask Refinement for Few-Shot Medical Image Segmentation
    :star:code
  • Graph-BAS3Net: Boundary-Aware Semi-Supervised Segmentation Network With Bilateral Graph Convolution医学图像分割
  • 病变分割
    • T-AutoML: Automated Machine Learning for Lesion Segmentation Using Transformers in 3D Medical Imaging
  • 息肉分割
    • Collaborative and Adversarial Learning of Focused and Dispersive Representations for Semi-Supervised Polyp Segmentation
  • 血管分割
    • Self-Supervised Vessel Segmentation via Adversarial Learning
  • 脑肿瘤分割
    • RFNet: Region-Aware Fusion Network for Incomplete Multi-Modal Brain Tumor Segmentation
• 病理学图像表示
  • A QuadTree Image Representation for Computational Pathology
• 医学图像分析
  • Preservational Learning Improves Self-supervised Medical Image Models by Reconstructing Diverse Contexts
    :star:code
    :newspaper:解读:ICCV2021 2D和3D通用！新医疗影像自监督SOTA（代码已开源）
• 医学图像去噪
  • Eformer: Edge Enhancement based Transformer for Medical Image Denoising
• 视频翻译
  • Long-Term Temporally Consistent Unpaired Video Translation From Simulated Surgical 3D Data
    :star:code:house:project
• 病理学图像核检测分割
  • Mutual-Complementing Framework for Nuclei Detection and Segmentation in Pathology Image
• 医学报告生成
  • Visual-Textual Attentive Semantic Consistency for Medical Report Generation
• CT
  • 3DeepCT: Learning Volumetric Scattering Tomography of Clouds
  • IntraTomo: Self-Supervised Learning-Based Tomography via Sinogram Synthesis and Prediction
    :star:code
  • CT重建
    • Dynamic CT Reconstruction From Limited Views With Implicit Neural Representations and Parametric Motion Fields
      :star:code
• 医学图像识别
  • GLoRIA: A Multimodal Global-Local Representation Learning Framework for Label-Efficient Medical Image Recognition
    :star:code
• 医学图像分类
  • Big Self-Supervised Models Advance Medical Image Classification
  • Large-Scale Robust Deep AUC Maximization: A New Surrogate Loss and Empirical Studies on Medical Image Classification
    :star:code

24.Face(人脸)

• VariTex: Variational Neural Face Textures
  :star:code:house:project:tv:video
• 人脸造假检测
  • OpenForensics: Large-Scale Challenging Dataset For Multi-Face Forgery Detection And Segmentation In-The-Wild
    :house:project
  • Exploring Temporal Coherence for More General Video Face Forgery Detection
• 人脸合成
  • Disentangled Lifespan Face Synthesis
    :star:code:house:project:tv:video
• 人脸识别
  • PASS: Protected Attribute Suppression System for Mitigating Bias in Face Recognition
  • SynFace: Face Recognition with Synthetic Data
  • Adaptive Label Noise Cleaning With Meta-Supervision for Deep Face Recognition
  • Disentangled Representation for Age-Invariant Face Recognition: A Mutual Information Minimization Perspective
  • Teacher-Student Adversarial Depth Hallucination To Improve Face Recognition
    :star:code
  • DAM: Discrepancy Alignment Metric for Face Recognition
  • “去”识别
    • Personalized and Invertible Face De-Identification by Disentangled Identity Information Manipulation
• Face perception面部感知
  • Learning Facial Representations from the Cycle-consistency of Face
    :star:code
• 说话人脸生成
  • FACIAL: Synthesizing Dynamic Talking Face with Implicit Attribute Learning
• 说话头合成
  • AD-NeRF: Audio Driven Neural Radiance Fields for Talking Head Synthesis
    :star:code
  • Learned Spatial Representations for Few-Shot Talking-Head Synthesis
    :star:code:house:project
• 人脸表情识别
  • Understanding and Mitigating Annotation Bias in Facial Expression Recognition
  • TransFER: Learning Relation-aware Facial Expression Representations with Transformers
• 人脸呈现攻击检测
  • Detection and Continual Learning of Novel Face Presentation Attacks
    :star:code
• 人脸编辑
  • Talk-to-Edit: Fine-Grained Facial Editing via Dialog
    :star:code:house:project
    :newspaper:解读:ICCV2021 | 南洋理工大学、港中大提出Talk-to-Edit，对话实现高细粒度人脸编辑
  • A Latent Transformer for Disentangled Face Editing in Images and Videos
    :star:code
• 人脸对齐
  • ADNet: Leveraging Error-Bias Towards Normal Direction in Face Alignment
• 人脸图像重建
  • Focal Frequency Loss for Image Reconstruction and Synthesis
    :star:code:house:project:tv:video
  • Towards High Fidelity Monocular Face Reconstruction With Rich Reflectance Using Self-Supervised Learning and Ray Tracing
  • Neural Photofit: Gaze-Based Mental Image Reconstruction
    :house:project
• 3D人脸重建
  • Topologically Consistent Multi-View Face Inference Using Volumetric Sampling
    :star:code
  • Self-Supervised 3D Face Reconstruction via Conditional Estimation
• 三维人脸动画
  • MeshTalk: 3D Face Animation From Speech Using Cross-Modality Disentanglement
    :star:code:tv:video
• Remote Photoplethysmography (rPPG远程光电容积描记术)
  • The Way to My Heart Is Through Contrastive Learning: Remote Photoplethysmography From Unlabelled Video
    :star:code
• 人脸加密
  • Towards Face Encryption by Generating Adversarial Identity Masks
    :star:code
• Deepfake检测
  • Learning Self-Consistency for Deepfake Detection
    :open_mouth:oral
  • Joint Audio-Visual Deepfake Detection
  • Artificial Fingerprinting for Generative Models: Rooting Deepfake Attribution in Training Data
    :open_mouth:oral
• 人脸纹理补全
  • Learning High-Fidelity Face Texture Completion Without Complete Face Texture
• 面部动作单元检测
  • PIAP-DF: Pixel-Interested and Anti Person-Specific Facial Action Unit Detection Net With Discrete Feedback Learning
• 人脸分析
  • Fake It Till You Make It: Face analysis in the wild using synthetic data alone
    :house:project:tv:video
• 3D头重建
  • H3D-Net: Few-Shot High-Fidelity 3D Head Reconstruction
• 人脸关键点检测
  • Improving Robustness of Facial Landmark Detection by Defending Against Adversarial Attacks
    :star:code
• 人脸图像检索
  • Face Image Retrieval with Attribute Manipulation

23.Gaze Estimation(视线估计)

• Generalizing Gaze Estimation with Outlier-guided Collaborative Adaptation
  :star:code
• 视线跟踪
  • Looking Here or There? Gaze Following in 360-Degree Images
• 视点估计
  • ViewNet: Unsupervised Viewpoint Estimation From Conditional Generation

22.GAN

• Sketch Your Own GAN
  :star:code:house:project
• Online Multi-Granularity Distillation for GAN Compression
  :star:code
• Dual Projection Generative Adversarial Networks for Conditional Image Generation
  :star:code
• InSeGAN: A Generative Approach to Segmenting Identical Instances in Depth Images
• ReStyle: A Residual-Based StyleGAN Encoder via Iterative Refinement
  :star:code:house:project:tv:video
• WarpedGANSpace: Finding non-linear RBF paths in GAN latent space
  :star:code
• Toward a Visual Concept Vocabulary for GAN Latent Space
• Collaging Class-specific GANs for Semantic Image Synthesis
  :house:project
• Latent Transformations via NeuralODEs for GAN-Based Image Editing
• Reality Transform Adversarial Generators for Image Splicing Forgery Detection and Localization
• GAN-Control: Explicitly Controllable GANs(https://alonshoshan10.github.io/gan_control/)
  :house:project
• Omni-GAN: On the Secrets of cGANs and Beyond
  :star:code
• Unsupervised Image Generation with Infinite Generative Adversarial Networks
  :star:code
• DAE-GAN: Dynamic Aspect-Aware GAN for Text-to-Image Synthesis
• Detail Me More: Improving GAN’s photo-realism of complex scenes
• Unsupervised Segmentation Incorporating Shape Prior via Generative Adversarial Networks
• DRB-GAN: A Dynamic ResBlock Generative Adversarial Network for Artistic Style Transfer
  :star:code
• Dual Contrastive Loss and Attention for GANs
• Semi-Supervised Single-Stage Controllable GANs for Conditional Fine-Grained Image Generation
• Gradient Normalization for Generative Adversarial Networks
  :star:code
• EigenGAN: Layer-Wise Eigen-Learning for GANs
  :star:code
• Retrieve in Style: Unsupervised Facial Feature Transfer and Retrieval
  :star:code
• HeadGAN: One-shot Neural Head Synthesis and Editing
  :house:project:tv:video
• Explaining in Style: Training a GAN To Explain a Classifier in StyleSpace
  :star:code:house:project:tv:video
• StyleCLIP: Text-Driven Manipulation of StyleGAN Imagery
  :open_mouth:oral:star:code:tv:video
• Towards Discovery and Attribution of Open-World GAN Generated Images
• Diagonal Attention and Style-Based GAN for Content-Style Disentanglement in Image Generation and Translation
• Re-Aging GAN: Toward Personalized Face Age Transformation
• When do GANs replicate? On the choice of dataset size
  :star:code
• LoFGAN: Fusing Local Representations for Few-shot Image Generation
• Multi-Class Multi-Instance Count Conditioned Adversarial Image Generation
  :star:code
• Generative Adversarial Registration for Improved Conditional Deformable Templates
  :star:code
• F-Drop&Match: GANs with a Dead Zone in the High-Frequency Domain
• GAN inversion(GAN逆映射)
  • From Continuity to Editability: Inverting GANs with Consecutive Images
    :star:code
  • GAN Inversion for Out-of-Range Images with Geometric Transformations
    :house:project
• 图像到图像翻译
  • Unaligned Image-to-Image Translation by Learning to Reweight
    :star:code
  • Bridging the Gap between Label- and Reference-based Synthesis in Multi-attribute Image-to-Image Translation
    :star:code
  • Instance-Wise Hard Negative Example Generation for Contrastive Learning in Unpaired Image-to-Image Translation
  • TransferI2I: Transfer Learning for Image-to-Image Translation from Small Datasets
    :star:code
  • Rethinking the Truly Unsupervised Image-to-Image Translation
  • SPatchGAN: A Statistical Feature Based Discriminator for Unsupervised Image-to-Image Translation
• Image translation
  • Scaling-up Disentanglement for Image Translation
    :star:code:house:project
  • Harnessing the Conditioning Sensorium for Improved Image Translation
  • Frequency Domain Image Translation: More Photo-Realistic, Better Identity-Preserving
    :star:code
  • Dual Transfer Learning for Event-based End-task Prediction via Pluggable Event to Image Translation
    :star:code
  • Semantically Robust Unpaired Image Translation for Data with Unmatched Semantics Statistics

21.Active Learning(主动学习)

• Semi-Supervised Active Learning with Temporal Output Discrepancy
  :star:code
• Influence Selection for Active Learning
• Active Domain Adaptation via Clustering Uncertainty-weighted Embeddings
  :star:code
• Contrastive Coding for Active Learning under Class Distribution Mismatch
  :star:code

20.Adversarial Learning(对抗学习)

• Low Curvature Activations Reduce Overfitting in Adversarial Training
• Removing Adversarial Noise in Class Activation Feature Space
• Sample Efficient Detection and Classification of Adversarial Attacks via Self-Supervised Embeddings
• Invisible Backdoor Attack With Sample-Specific Triggers
  :star:code
• Defending Against Universal Adversarial Patches by Clipping Feature Norms
• 对抗攻击
  • Feature Importance-aware Transferable Adversarial Attacks
    :star:code
  • TkML-AP: Adversarial Attacks to Top-k Multi-Label Learning
    :star:code
  • Meta Gradient Adversarial Attack
  • AGKD-BML: Defense Against Adversarial Attack by Attention Guided Knowledge Distillation and Bi-directional Metric Learning
    :star:code
  • Exploiting Multi-Object Relationships for Detecting Adversarial Attacks in Complex Scenes
  • AdvDrop: Adversarial Attack to DNNs by Dropping Information
    :star:code
  • Adversarial Attacks Are Reversible With Natural Supervision
  • Attack As the Best Defense: Nullifying Image-to-Image Translation GANs via Limit-Aware Adversarial Attack
  • Learnable Boundary Guided Adversarial Training
    :star:code
  • Augmented Lagrangian Adversarial Attacks
    :star:code
  • Meta-Attack: Class-Agnostic and Model-Agnostic Physical Adversarial Attack
  • On Generating Transferable Targeted Perturbations
    :star:code
  • Admix: Enhancing the Transferability of Adversarial Attacks
    :star:code
  • Consistency-Sensitivity Guided Ensemble Black-Box Adversarial Attacks in Low-Dimensional Spaces
  • Adversarial Attacks On Multi-Agent Communication
  • Interpreting Attributions and Interactions of Adversarial Attacks
  • RDA: Robust Domain Adaptation via Fourier Adversarial Attacking
• 对抗样本
  • Adversarial Example Detection Using Latent Neighborhood Graph
  • On the Robustness of Vision Transformers to Adversarial Examples
• 黑盒
  • Black-Box Detection of Backdoor Attacks With Limited Information and Data
  • Aha! Adaptive History-Driven Attack for Decision-Based Black-Box Models
  • Data-Free Universal Adversarial Perturbation and Black-Box Attack

19.Self-Driving Vehicles(自动驾驶)

• End-to-End Urban Driving by Imitating a Reinforcement Learning Coach
  :star:code
• MultiSiam: Self-supervised Multi-instance Siamese Representation Learning for Autonomous Driving
  :star:code
• NEAT: Neural Attention Fields for End-to-End Autonomous Driving
  :star:code
• Safety-aware Motion Prediction with Unseen Vehicles for Autonomous Driving
  :star:code
• Social-NCE: Contrastive Learning of Socially-aware Motion Representations
  :star:code:tv:video
• Learning To Drive From a World on Rails
  :open_mouth:oral:star:code:house:project
• DRIVE: Deep Reinforced Accident Anticipation With Visual Explanation
  :star:code:house:project:tv:video
• LookOut: Diverse Multi-Future Prediction and Planning for Self-Driving
• Prediction by Anticipation: An Action-Conditional Prediction Method Based on Interaction Learning
  :star:code:tv:video
• TMCOSS: Thresholded Multi-Criteria Online Subset Selection for Data-Efficient Autonomous Driving
• FIERY: Future Instance Prediction in Bird's-Eye View From Surround Monocular Cameras
  :star:code
• On Exposing the Challenging Long Tail in Future Prediction of Traffic Actors
  :star:code
• MGNet: Monocular Geometric Scene Understanding for Autonomous Driving
  :star:code:tv:video
• Human trajectory prediction(人体轨迹预测)
  • Personalized Trajectory Prediction via Distribution Discrimination
    :star:code
  • Human Trajectory Prediction via Counterfactual Analysis
    :star:code
  • From Goals, Waypoints & Paths to Long Term Human Trajectory Forecasting
    :star:code:house:project:tv:video
• 轨迹预测
  • Unlimited Neighborhood Interaction for Heterogeneous Trajectory Prediction
  • LOKI: Long Term and Key Intentions for Trajectory Prediction
  • MG-GAN: A Multi-Generator Model Preventing Out-of-Distribution Samples in Pedestrian Trajectory Prediction
    :star:code
  • DenseTNT: End-to-end Trajectory Prediction from Dense Goal Sets
  • Where Are You Heading? Dynamic Trajectory Prediction With Expert Goal Examples
    :star:code
  • Three Steps to Multimodal Trajectory Prediction: Modality Clustering, Classification and Synthesis
  • Spatial-Temporal Consistency Network for Low-Latency Trajectory Forecasting
  • Likelihood-Based Diverse Sampling for Trajectory Forecasting
    :star:code
• 运动预测
  • RAIN: Reinforced Hybrid Attention Inference Network for Motion Forecasting
    :house:project
  • SLAMP: Stochastic Latent Appearance and Motion Prediction
    :house:project
  • SlowFast Rolling-Unrolling LSTMs for Action Anticipation in Egocentric Videos
• 自动导航
  • FOVEA: Foveated Image Magnification for Autonomous Navigation
    :house:project
• 交通场景理解
  • Structured Bird's-Eye-View Traffic Scene Understanding from Onboard Images
    :star:code
• 车辆车牌识别
  • 车辆重识别
    • Heterogeneous Relational Complement for Vehicle Re-identification
    • Self-Supervised Geometric Features Discovery via Interpretable Attention for Vehicle Re-Identification and Beyond
• 自主赛车
  • Learn-to-Race: A Multimodal Control Environment for Autonomous Racing
    :star:code
• 预测司机的视觉注意力
  • MEDIRL: Predicting the Visual Attention of Drivers via Maximum Entropy Deep Inverse Reinforcement Learning
    :star:code
• 姿势预测
  • Space-Time-Separable Graph Convolutional Network for Pose Forecasting
    :star:code:tv:video
• 车辆跟踪
  • Track Without Appearance: Learn Box and Tracklet Embedding With Local and Global Motion Patterns for Vehicle Tracking
• 对任意相机视角的车辆进行检测分析
  • Robust 2D/3D Vehicle Parsing in Arbitrary Camera Views for CVIS
• 车道线检测
  • CondLaneNet: A Top-To-Down Lane Detection Framework Based on Conditional Convolution
    :star:code
  • Active Learning for Lane Detection: A Knowledge Distillation Approach
• 车速估计
  • Robust Automatic Monocular Vehicle Speed Estimation for Traffic Surveillance
    :star:code

18.Transformers

• Generic Attention-model Explainability for Interpreting Bi-Modal and Encoder-Decoder Transformers
  :open_mouth:oral:star:code
  :newspaper:解读:ICCV2021 Oral-TAU&Facebook提出了通用的Attention模型可解释性
• Transformer-Based Attention Networks for Continuous Pixel-Wise Prediction
  :star:code
• PlaneTR: Structure-Guided Transformers for 3D Plane Recovery
  :star:code
• Rethinking and Improving Relative Position Encoding for Vision Transformer
  :star:code
• Vision Transformer with Progressive Sampling
  :star:code
• Paint Transformer: Feed Forward Neural Painting with Stroke Prediction
  :open_mouth:oral:star:code
• Rethinking Spatial Dimensions of Vision Transformers
  :star:code
  :newspaper:解读:ICCV2021-PiT-池化操作不是CNN的专属，ViT说：“我也可以”；南大提出池化视觉Transformer（PiT）
• PnP-DETR: Towards Efficient Visual Analysis with Transformers
  :star:code
• Describing and Localizing Multiple Changes With Transformers
  :star:code:house:project
• LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference
  :star:code
• VidTr: Video Transformer Without Convolutions
• Visformer: The Vision-Friendly Transformer
  :star:code
• Going Deeper With Image Transformers
  :star:code
• Multiscale Vision Transformers
  :star:code
• Learning Multi-Scene Absolute Pose Regression With Transformers
  :star:code
• Visual Saliency Transformer
  :star:code
• Event-Based Video Reconstruction Using Transformer
  :star:code
• Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows
  :star:code
• An Empirical Study of Training Self-Supervised Vision Transformers
  :open_mouth:oral:star:code
• Tokens-to-Token ViT: Training Vision Transformers From Scratch on ImageNet
  :star:code
• CvT: Introducing Convolutions to Vision Transformers
  :star:code
• COTR: Correspondence Transformer for Matching Across Images
• ViViT: A Video Vision Transformer
  :star:code
• AgentFormer: Agent-Aware Transformers for Socio-Temporal Multi-Agent Forecasting
  :star:code:house:project
• Incorporating Convolution Designs into Visual Transformers
  :star:code
• LayoutTransformer: Layout Generation and Completion with Self-attention
  :star:code:house:project
• AutoFormer: Searching Transformers for Visual Recognition
  :star:code
• Scalable Vision Transformers With Hierarchical Pooling
  :star:code
• Visual Transformers: Where Do Transformers Really Belong in Vision Models?
• Anticipative Video Transformer
  :star:code:house:project
• 密集预测
  • Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions
    :open_mouth:oral:star:code
    :newspaper:解读:大白话Pyramid Vision Transformer
  • Vision Transformers for Dense Prediction
    :star:code
• 3D人体纹理估计
  • 3D Human Texture Estimation from a Single Image with Transformers
    :open_mouth:oral:house:project
• 图像编辑
  • Multi-Scale Vision Longformer: A New Vision Transformer for High-Resolution Image Encoding
    :star:code
• OCR
  • DocFormer: End-to-End Transformer for Document Understanding
• 根据音乐生成舞蹈
  • AI Choreographer: Music Conditioned 3D Dance Generation With AIST++
    :house:project
    :newspaper:简介:Transformer又又来了，生成配有音乐的丝滑3D舞蹈，开放最大规模数据集AIST++

17.3D(三维视觉)

• Discovering 3D Parts from Image Collections
  :open_mouth:oral:star:code:house:project:tv:video
• PixelSynth: Generating a 3D-Consistent Experience from a Single Image
  :star:code:house:project
• Towers of Babel: Combining Images, Language, and 3D Geometry for Learning Multimodal Vision
  :star:code:house:project
• Pixel-Perfect Structure-from-Motion with Featuremetric Refinement
  :star:code
• Learning Anchored Unsigned Distance Functions with Gradient Direction Alignment for Single-view Garment Reconstruction
  :open_mouth:oral:star:code
• LSD-StructureNet: Modeling Levels of Structural Detail in 3D Part Hierarchies
• Rational Polynomial Camera Model Warping for Deep Learning Based Satellite Multi-View Stereo Matching
  :star:code
• Where2Act: From Pixels to Actions for Articulated 3D Objects
  :tv:video
• BuildingNet: Learning to Label 3D Buildings
  :open_mouth:oral:star:code:house:project
• SurfGen: Adversarial 3D Shape Synthesis With Explicit Surface Discriminators
• Deep Virtual Markers for Articulated 3D Shapes
  :star:code:tv:video
• Learning Efficient Photometric Feature Transform for Multi-view Stereo
  :house:project
• Neural-GIF: Neural Generalized Implicit Functions for Animating People in Clothing
• Just a Few Points Are All You Need for Multi-View Stereo: A Novel Semi-Supervised Learning Method for Multi-View Stereo
• 3D-FRONT: 3D Furnished Rooms With layOuts and semaNTics
• Learning Generative Models of Textured 3D Meshes from Real-World Images
  :star:code
• Self-Supervised Pretraining of 3D Features on any Point-Cloud
• High Quality Disparity Remapping with Two-Stage Warping
• Structure-From-Sherds: Incremental 3D Reassembly of Axially Symmetric Pots From Unordered and Mixed Fragment Collections
  :star:code
• Interpolation-Aware Padding for 3D Sparse Convolutional Neural Networks
• 深度估计
  • StructDepth: Leveraging the structural regularities for self-supervised indoor depth estimation
    :star:code
  • Bridging Unsupervised and Supervised Depth from Focus via All-in-Focus Supervision
    :star:code:house:project
  • Augmenting Depth Estimation with Geospatial Context
  • Can Scale-Consistent Monocular Depth Be Learned in a Self-Supervised Scale-Invariant Manner?
  • Revisiting Stereo Depth Estimation From a Sequence-to-Sequence Perspective With Transformers
    :open_mouth:oral:star:code
  • Adaptive Surface Normal Constraint for Depth Estimation
    :star:code
  • Event-Intensity Stereo: Estimating Depth by the Best of Both Worlds
  • DnD: Dense Depth Estimation in Crowded Dynamic Indoor Scenes
  • DepthInSpace: Exploitation and Fusion of Multiple Video Frames for Structured-Light Depth Estimation
    :house:project
  • Boosting Monocular Depth Estimation With Lightweight 3D Point Fusion
  • Monocular Depth Estimation(单目深度估计)
    • Revealing the Reciprocal Relations Between Self-Supervised Stereo and Monocular Depth Estimation
    • MonoIndoor: Towards Good Practice of Self-Supervised Monocular Depth Estimation for Indoor Environments
    • Regularizing Nighttime Weirdness: Efficient Self-supervised Monocular Depth Estimation in the Dark
      :star:code
    • Towards Interpretable Deep Networks for Monocular Depth Estimation
      :star:code:tv:video
    • Self-supervised Monocular Depth Estimation for All Day Images using Domain Separation
      :star:code
    • Fine-grained Semantics-aware Representation Enhancement for Self-supervised Monocular Depth Estimation
      :open_mouth:oral:star:code
    • Excavating the Potential Capacity of Self-Supervised Monocular Depth Estimation
      :star:code
    • R-MSFM: Recurrent Multi-Scale Feature Modulation for Monocular Depth Estimating
      :star:code
    • Adaptive Confidence Thresholding for Monocular Depth Estimation(https://github.com/megvii-research/OMNet)
    • SaccadeCam: Adaptive Visual Attention for Monocular Depth Sensing
• 深度补全
  • Bayesian Deep Basis Fitting for Depth Completion With Uncertainty
  • Unsupervised Depth Completion With Calibrated Backprojection Layers
    :star:code
• Omnidirectional Localization
  • PICCOLO: Point Cloud-Centric Omnidirectional Localization
• 三维重建
  • Learning Signed Distance Field for Multi-view Surface Reconstruction
    :open_mouth:oral
  • 3DStyleNet: Creating 3D Shapes with Geometric and Texture Style Variations
    :open_mouth:oral:house:project
  • DensePose 3D: Lifting Canonical Surface Maps of Articulated Objects to the Third Dimension
  • In-the-Wild Single Camera 3D Reconstruction Through Moving Water Surfaces
    :open_mouth:oral:star:code:tv:video
  • Gaussian Fusion: Accurate 3D Reconstruction via Geometry-Guided Displacement Interpolation
  • RetrievalFuse: Neural 3D Scene Reconstruction With a Database
    :star:code:house:project:tv:video
  • Multi-View 3D Reconstruction With Transformers
  • Polarimetric Helmholtz Stereopsis
  • MINE: Towards Continuous Depth MPI with NeRF for Novel View Synthesis
    :star:code:tv:video
  • Toward Realistic Single-View 3D Object Reconstruction With Unsupervised Learning From Multiple Images
    :star:code
  • CryoDRGN2: Ab initio neural reconstruction of 3D protein structures from real cryo-EM images
  • 三维场景重建
    • Graph-to-3D: End-to-End Generation and Manipulation of 3D Scenes Using Scene Graphs
    • VolumeFusion: Deep Depth Fusion for 3D Scene Reconstruction
  • 三维形状重建
    • 3DIAS: 3D Shape Reconstruction with Implicit Algebraic Surfaces
      :house:project
    • Multiresolution Deep Implicit Functions for 3D Shape Representation
    • MVTN: Multi-View Transformation Network for 3D Shape Recognition
      :star:code:tv:video
    • Sketch2Mesh: Reconstructing and Editing 3D Shapes from Sketches
  • 三维网格重建
    • Vis2Mesh: Efficient Mesh Reconstruction from Unstructured Point Clouds of Large Scenes with Learned Virtual View Visibility
      :star:code
• 三维场景
  • Scene Synthesis via Uncertainty-Driven Attribute Synchronization
    :star:code
  • Graph-to-3D: End-to-End Generation and Manipulation of 3D Scenes Using Scene Graphs
    :house:project
• 相机校准
  • CTRL-C: Camera calibration TRansformer with Line-Classification
    :star:code
  • BabelCalib: A Universal Approach to Calibrating Central Cameras
    :star:code
• 表面重建
  • Temporally-Coherent Surface Reconstruction via Metric-Consistent Atlases
  • Planar Surface Reconstruction from Sparse Views
    :open_mouth:oral:star:code:house:project:tv:video
  • Adaptive Surface Reconstruction With Multiscale Convolutional Kernels
• 3D场景合成
  • Indoor Scene Generation From a Collection of Semantic-Segmented Depth Images
    :star:code
• 3D形状识别
  • Learning Canonical View Representation for 3D Shape Recognition With Arbitrary Views
    :star:code
• 图像重建
  • Semantic-embedded Unsupervised Spectral Reconstruction from Single RGB Images in the Wild
    :star:code
• Multi-view Stereo(MVS)
  • Digging into Uncertainty in Self-supervised Multi-view Stereo
    :star:code
  • PatchMatch-RL: Deep MVS with Pixelwise Depth, Normal, and Visibility
    :star:code
  • A Confidence-Based Iterative Solver of Depths and Surface Normals for Deep Multi-View Stereo
    :star:code
  • EPP-MVSNet: Epipolar-Assembling Based Depth Prediction for Multi-View Stereo

16.Re-Identification(重识别)

Object Re-Identification目标(物体)重识别

• TransReID: Transformer-Based Object Re-Identification
  :star:code

Person Re-Identification(人员重识别)

• Spatio-Temporal Representation Factorization for Video-based Person Re-Identification
• Learning Instance-level Spatial-Temporal Patterns for Person Re-identification
  :star:code
• Towards Discriminative Representation Learning for Unsupervised Person Re-identification
• Learning by Aligning: Visible-Infrared Person Re-identification using Cross-Modal Correspondences
  :star:code:house:project
• Video-based Person Re-identification with Spatial and Temporal Memory Networks
  :house:project
• Multi-Expert Adversarial Attack Detection in Person Re-identification Using Context Inconsistency
• Clothing Status Awareness for Long-Term Person Re-Identification
• Dense Interaction Learning for Video-Based Person Re-Identification
  :open_mouth:oral
• Explainable Person Re-Identification With Attribute-Guided Metric Distillation
  :house:project
• Online Pseudo Label Generation by Hierarchical Cluster Dynamics for Adaptive Person Re-Identification
• Pyramid Spatial-Temporal Aggregation for Video-Based Person Re-Identification
  :star:code
• ICE: Inter-Instance Contrastive Encoding for Unsupervised Person Re-Identification
  :star:code:tv:video
• Learning To Know Where To See: A Visibility-Aware Approach for Occluded Person Re-Identification
• Attack-Guided Perceptual Data Generation for Real-world Re-Identification
• BV-Person: A Large-Scale Dataset for Bird-View Person Re-Identification
  :sunflower:dataset
• CM-NAS: Cross-Modality Neural Architecture Search for Visible-Infrared Person Re-Identification
  :star:code
• Meta Pairwise Relationship Distillation for Unsupervised Person Re-Identification
• Syncretic Modality Collaborative Learning for Visible Infrared Person Re-Identification
• Weakly Supervised Text-Based Person Re-Identification
  :star:code
• Occlude Them All: Occlusion-Aware Attention Network for Occluded Person Re-ID
• Occluded Person Re-Identification with Single-scale Global Representations
  :star:code
• 域适应人员重识别
  • IDM: An Intermediate Domain Module for Domain Adaptive Person Re-ID
    :open_mouth:oral:star:code
• Crowd Counting(拥挤人群计数)
  • Rethinking Counting and Localization in Crowds:A Purely Point-Based Framework
    :open_mouth:oral:star:code
  • Uniformity in Heterogeneity:Diving Deep into Count Interval Partition for Crowd Counting
    :star:code
  • Spatial Uncertainty-Aware Semi-Supervised Crowd Counting
    :star:code
  • Variational Attention: Propagating Domain-Specific Knowledge for Multi-Domain Learning in Crowd Counting
    :star:code
  • Exploiting Sample Correlation for Crowd Counting With Multi-Expert Network
    :star:code
  • Crowd Counting With Partial Annotations in an Image
    :star:code
  • Towards A Universal Model for Cross-Dataset Crowd Counting
  • Rethinking Counting and Localization in Crowds:A Purely Point-Based Framework
    :open_mouth:oral:star:code
  • Uniformity in Heterogeneity: Diving Deep Into Count Interval Partition for Crowd Counting
    :star:code
• 跨模态人员重识别
  • Cross-Modality Person Re-Identification via Modality Confusion and Center Aggregation
• 行人检测
  • MOTSynth: How Can Synthetic Data Help Pedestrian Detection and Tracking?
  • Stacked Homography Transformations for Multi-View Pedestrian Detection
  • Robust Small-scale Pedestrian Detection with Cued Recall via Memory Learning
  • Body-Face Joint Detection via Embedding and Head Hook
    :star:code
• 行人属性识别
  • Spatial and Semantic Consistency Regularizations for Pedestrian Attribute Recognition
  • LapsCore: Language-Guided Person Search via Color Reasoning
• Person Search(行人搜索)
  • Weakly Supervised Person Search with Region Siamese Networks
  • End-to-End Trainable Trident Person Search Network Using Adaptive Gradient Propagation
  • ASMR: Learning Attribute-Based Person Search with Adaptive Semantic Margin Regularizer
    :house:project
  • Weakly Supervised Person Search with Region Siamese Networks
• 行人行为预测
  • Bifold and Semantic Reasoning for Pedestrian Behavior Prediction
• 步态识别
  • Gait Recognition via Effective Global-Local Feature Representation and Local Temporal Aggregation
  • Context-Sensitive Temporal Feature Learning for Gait Recognition
    :star:code
  • 3D Local Convolutional Neural Networks for Gait Recognition
    :star:code
  • Gait Recognition in the Wild: A Benchmark
    :star:code:house:project

15.Object Tracking(目标跟踪)

• Saliency-Associated Object Tracking
• Box-Aware Feature Enhancement for Single Object Tracking on Point Clouds
  :star:code
• Learning to Track Objects from Unlabeled Videos
  :star:code
• DepthTrack : Unveiling the Power of RGBD Tracking
  :star:code
• Learning Target Candidate Association To Keep Track of What Not To Track
  :star:code
• Transparent Object Tracking Benchmark
  :house:project
• DepthTrack: Unveiling the Power of RGBD Tracking
  :star:code
• Object Tracking by Jointly Exploiting Frame and Event Domain
• High-Performance Discriminative Tracking with Transformers
• Visio-Temporal Attention for Multi-Camera Multi-Target Association
• 视觉目标跟踪
  • Learning to Adversarially Blur Visual Object Tracking
    :star:code
  • Learn to Match: Automatic Matching Network Design for Visual Tracking
    :star:code
  • Video Annotation for Visual Tracking via Selection and Refinement
    :star:code
  • Learning Spatio-Temporal Transformer for Visual Tracking
    :star:code
  • 3D视觉跟踪
    • MLVSNet: Multi-Level Voting Siamese Network for 3D Visual Tracking
      :star:code
• 卫星图像跟踪
  • HiFT: Hierarchical Feature Transformer for Aerial Tracking
    :star:code
• 3D多目标跟踪
  • Exploring Simple 3D Multi-Object Tracking for Autonomous Driving
    :star:code
    :newspaper:解读:ICCV 2021丨轻舟智航提出SimTrack: 3D多目标一体化检测与跟踪，简单又精确
• 多目标跟踪与分割
  • Assignment-Space-Based Multi-Object Tracking and Segmentation
  • Continuous Copy-Paste for One-Stage Multi-Object Tracking and Segmentation
    :star:code
• 多目标跟踪
  • A General Recurrent Tracking Framework Without Real Data
  • Making Higher Order MOT Scalable: An Efficient Approximate Solver for Lifted Disjoint Paths
    :star:code
  • MOTSynth: How Can Synthetic Data Help Pedestrian Detection and Tracking?
  • Learning To Track With Object Permanence
    :star:code多目标跟踪
• 视频目标跟踪
  • TF-Blender: Temporal Feature Blender for Video Object Detection
    :star:code

14.Object Detection(目标检测)

• Rank & Sort Loss for Object Detection and Instance Segmentation
  :open_mouth:oral:star:code
• MDETR : Modulated Detection for End-to-End Multi-Modal Understanding
  :open_mouth:oral:star:code
• SimROD: A Simple Adaptation Method for Robust Object Detection
  :open_mouth:oral:house:project
  :newspaper:解读:ICCV2021 Oral SimROD：简单高效的数据增强！华为提出了一种简单的鲁棒目标检测自适应方法
• GraphFPN: Graph Feature Pyramid Network for Object Detection
• Fast Convergence of DETR with Spatially Modulated Co-Attention
  :star:code
• Oriented R-CNN for Object Detection
  :star:code
• Conditional DETR for Fast Training Convergence
  :newspaper:解读:通过显式寻找物体的 extremity 区域加快 DETR 的收敛：Conditional DETR
• Vector-Decomposed Disentanglement for Domain-Invariant Object Detection
  :star:code
• G-DetKD: Towards General Distillation Framework for Object Detectors via Contrastive and Semantic-guided Feature Imitation
• ODAM: Object Detection, Association, and Mapping using Posed RGB Video
  :open_mouth:oral
• Reconcile Prediction Consistency for Balanced Object Detection
• Deep Structured Instance Graph for Distilling Object Detectors
  :star:code
• Towards Rotation Invariance in Object Detection
  :star:code
• Morphable Detector for Object Detection on Demand
  :star:code
• DetCo: Unsupervised Contrastive Learning for Object Detection
  :star:code
• Domain-Invariant Disentangled Network for Generalizable Object Detection
• MDETR - Modulated Detection for End-to-End Multi-Modal Understanding
  :star:code
• Detecting Persuasive Atypicality by Modeling Contextual Compatibility
  :star:code
• Wanderlust: Online Continual Object Detection in the Real World
  :house:project
• PreDet: Large-Scale Weakly Supervised Pre-Training for Detection
• FMODetect: Robust Detection of Fast Moving Objects
• Multi-Source Domain Adaptation for Object Detection
• Self-Supervised Object Detection via Generative Image Synthesis
  :star:code
• Naturalistic Physical Adversarial Patch for Object Detectors
  :star:code
• Rethinking Transformer-Based Set Prediction for Object Detection
  :star:code
• Detecting Invisible People
  :house:project:tv:video
• Dynamic DETR: End-to-End Object Detection With Dynamic Attention
• CrossDet: Crossline Representation for Object Detection
  :star:code
• Robust Object Detection via Instance-Level Temporal Cycle Confusion
  :star:code
• End-to-End Semi-Supervised Object Detection With Soft Teacher
  :star:code
• Parallel Rectangle Flip Attack: A Query-Based Black-Box Attack Against Object Detection
• Fooling LiDAR Perception via Adversarial Trajectory Perturbation
  :star:code:house:project
• TOOD: Task-Aligned One-Stage Object Detection
  :open_mouth:oral:star:code
• Active Learning for Deep Object Detection via Probabilistic Modeling
  :star:code
  :newspaper:解读:ICCV2021 还在用大量数据暴力train模型？主动学习，教你选出数据集中最有价值的样本
• Dual Bipartite Graph Learning: A General Approach for Domain Adaptive Object Detection
• WB-DETR: Transformer-Based Detector without Backbone
• 3D目标检测
  • Geometry Uncertainty Projection Network for Monocular 3D Object Detection
    :star:code
  • Fog Simulation on Real LiDAR Point Clouds for 3D Object Detection in Adverse Weather
    :star:code:house:project
  • Is Pseudo-Lidar needed for Monocular 3D Object detection?
    :star:code
  • RandomRooms: Unsupervised Pre-training from Synthetic Shapes and Randomized Layouts for 3D Object Detection
  • LIGA-Stereo: Learning LiDAR Geometry Aware Representations for Stereo-based 3D Detector
    :star:code:house:project
  • Improving 3D Object Detection with Channel-wise Transformer
  • 4D-Net for Learned Multi-Modal Alignment
  • Pyramid R-CNN: Towards Better Performance and Adaptability for 3D Object Detection
  • Voxel Transformer for 3D Object Detection
  • An End-to-End Transformer Model for 3D Object Detection
    :open_mouth:oral:star:code:house:project
  • Unsupervised Domain Adaptive 3D Detection With Multi-Level Consistency
  • Group-Free 3D Object Detection via Transformers
    :star:code
  • VENet: Voting Enhancement Network for 3D Object Detection
  • Multi-Echo LiDAR for 3D Object Detection
  • Voxel Transformer for 3D Object Detection
  • RangeDet: In Defense of Range View for LiDAR-Based 3D Object Detection
    :star:code
  • The Devil Is in the Task: Exploiting Reciprocal Appearance-Localization Features for Monocular 3D Object Detection
  • Gated3D: Monocular 3D Object Detection From Temporal Illumination Cues
    :house:project
  • Are We Missing Confidence in Pseudo-LiDAR Methods for Monocular 3D Object Detection?
  • SPG: Unsupervised Domain Adaptation for 3D Object Detection via Semantic Point Generation
  • You Don't Only Look Once: Constructing Spatial-Temporal Memory for Integrated 3D Object Detection and Tracking
    :star:code:house:project:tv:video
  • Exploring Geometry-Aware Contrast and Clustering Harmonization for Self-Supervised 3D Object Detection
  • AutoShape: Real-Time Shape-Aware Monocular 3D Object Detection
    :star:code
  • Geometry-Based Distance Decomposition for Monocular 3D Object Detection
    :star:code
• 目标定位
  • Contrastive Attention Maps for Self-supervised Co-localization
  • TS-CAM: Token Semantic Coupled Attention Map for Weakly Supervised Object Localization
    :star:code
  • 弱监督目标定位
    • Normalization Matters in Weakly Supervised Object Localization
    • Online Refinement of Low-Level Feature Based Activation Map for Weakly Supervised Object Localization
    • Foreground Activation Maps for Weakly Supervised Object Localization
• Anomaly Detection(图像异常检测)
  • Divide-and-Assemble: Learning Block-wise Memory for Unsupervised Anomaly Detection
• 弱监督目标检测
  • Boosting Weakly Supervised Object Detection via Learning Bounding Box Adjusters
    :star:code
  • CaT: Weakly Supervised Object Detection With Category Transfer
    :star:code
• OOD 检测
  • Triggering Failures: Out-Of-Distribution detection by learning from local adversarial attacks in Semantic Segmentation
    :star:code
• 显著目标检测
  • Disentangled High Quality Salient Object Detection
  • Specificity-preserving RGB-D Saliency Detection
    :star:code
  • Light Field Saliency Detection with Dual Local Graph Learning and Reciprocative Guidance
  • MFNet: Multi-Filter Directive Network for Weakly Supervised Salient Object Detection
    :star:code
  • Scene Context-Aware Salient Object Detection
    :star:code
  • Dynamic Context-Sensitive Filtering Network for Video Salient Object Detection
    :star:code
  • iNAS: Integral NAS for Device-Aware Salient Object Detection
    :house:project
  • RGB-D显著目标检测
    • RGB-D Saliency Detection via Cascaded Mutual Information Minimization
      :star:code
  • co-saliency detection
    • Summarize and Search: Learning Consensus-aware Dynamic Convolution for Co-Saliency Detection
      :star:code
• 违禁物品检测
  • Towards Real-World Prohibited Item Detection: A Large-Scale X-ray Benchmark
    :sunflower:dataset
  • Towards Real-World X-Ray Security Inspection: A High-Quality Benchmark and Lateral Inhibition Module for Prohibited Items Detection
    :star:code
• 小样本目标检测
  • DeFRCN: Decoupled Faster R-CNN for Few-Shot Object Detection
    :star:code
  • Query Adaptive Few-Shot Object Detection With Heterogeneous Graph Convolutional Networks
  • Universal-Prototype Enhancing for Few-Shot Object Detection
    :star:code
• 视觉关系协同定位
  • Few-shot Visual Relationship Co-localization
    :star:code:house:project
• 密集目标检测
  • Mutual Supervision for Dense Object Detection
    :star:code
• 域适应目标检测
  • Seeking Similarities over Differences: Similarity-based Domain Alignment for Adaptive Object Detection
  • Knowledge Mining and Transferring for Domain Adaptive Object Detection
• 图像篡改检测
  • Image Manipulation Detection by Multi-View Multi-Scale Supervision
    :star:code
• Visual Relationship Detection(VRD视觉关系检测)
  • Grounding Consistency: Distilling Spatial Common Sense for Precise Visual Relationship Detection
• 长尾目标检测
  • Exploring Classification Equilibrium in Long-Tailed Object Detection
    :star:code
  • MosaicOS: A Simple and Effective Use of Object-Centric Images for Long-Tailed Object Detection
    :star:code
• Salient Object Ranking
  • Salient Object Ranking with Position-Preserved Attention
    :star:code
• 小目标检测
  • Robust Small Object Detection on the Water Surface Through Fusion of Camera and Millimeter Wave Radar
• 黑暗中目标检测
  • Multitask AET With Orthogonal Tangent Regularity for Dark Object Detection
    :star:code
• 3D object prediction
  • Holistic Pose Graph: Modeling Geometric Structure among Objects in a Scene using Graph Inference for 3D Object Prediction
    :star:code
• 多目标检测
  • Training Multi-Object Detector by Estimating Bounding Box Distribution for Input Image
    :star:code
• 3D object grounding
  • Free-form Description Guided 3D Visual Graph Network for Object Grounding in Point Cloud
    :star:code
• 细粒度裂纹检测
  • CrackFormer: Transformer Network for Fine-Grained Crack Detection
• 线段检测
  • ELSD: Efficient Line Segment Detector and Descriptor
• 细胞检测与分类
  • Multi-Class Cell Detection Using Spatial Context Representation
    :star:code
• 阴影检测
  • Mitigating Intensity Bias in Shadow Detection via Feature Decomposition and Reweighting
• 社交距离检测
  • BEV-Net: Assessing Social Distancing Compliance by Joint People Localization and Geometric Reasoning
    :star:code
• 伪装目标检测
  • Uncertainty-Guided Transformer Reasoning for Camouflaged Object Detection
    :star:code

13.Image Segmentation(图像分割)

• Standardized Max Logits: A Simple yet Effective Approach for Identifying Unexpected Road Obstacles in Urban-Scene Segmentation
  :open_mouth:oral:star:code:tv:video
• TransForensics: Image Forgery Localization with Dense Self-Attention
• From Contexts to Locality: Ultra-high Resolution Image Segmentation via Locality-aware Contextual Correlation
  :star:code
• Labels4Free: Unsupervised Segmentation using StyleGAN
  :house:project:tv:video
• Warp-Refine Propagation: Semi-Supervised Auto-labeling via Cycle-consistency
• Scaling up instance annotation via label propagation
  :star:code:house:project
• Robust Trust Region for Weakly Supervised Segmentation
  :star:code:tv:video
• HPNet: Deep Primitive Segmentation Using Hybrid Representations
  :star:code
• Weakly Supervised Segmentation of Small Buildings With Point Labels
• BAPA-Net: Boundary Adaptation and Prototype Alignment for Cross-Domain Semantic Segmentation
  :star:code
• Conditional Diffusion for Interactive Segmentation
• Human Detection and Segmentation via Multi-view Consensus
  :star:code
• Unidentified Video Objects: A Benchmark for Dense, Open-World Segmentation
• Enhanced Boundary Learning for Glass-Like Object Segmentation
  :star:code
• PARTS: Unsupervised segmentation with slots, attention and independence maximization
• Predictive Feature Learning for Future Segmentation Prediction
• Perception-Aware Multi-Sensor Fusion for 3D LiDAR Semantic Segmentation
  :star:code
• Segmenter: Transformer for Semantic Segmentation
  :star:code
• C3-SemiSeg: Contrastive Semi-Supervised Segmentation via Cross-Set Learning and Dynamic Class-Balancing
• 全景分割
  • Panoptic Narrative Grounding
• 语义分割
  • Re-distributing Biased Pseudo Labels for Semi-supervised Semantic Segmentation: A Baseline Investigation
    :open_mouth:oral:star:code
  • Personalized Image Semantic Segmentation
    :star:code
  • RECALL: Replay-based Continual Learning in Semantic Segmentation
    :star:code
  • Deep Metric Learning for Open World Semantic Segmentation
  • LabOR: Labeling Only if Required for Domain Adaptive Semantic Segmentation
    :open_mouth:oral
  • Dual Path Learning for Domain Adaptation of Semantic Segmentation
    :star:code
  • Multi-Target Adversarial Frameworks for Domain Adaptation in Semantic Segmentation
  • Exploiting a Joint Embedding Space for Generalized Zero-Shot Semantic Segmentation
    :star:code:house:project
  • Multi-Anchor Active Domain Adaptation for Semantic Segmentation
    :open_mouth:oral:star:code
  • Pixel Contrastive-Consistent Semi-Supervised Semantic Segmentation
  • Self-Regulation for Semantic Segmentation
    :star:code
  • ShapeConv: Shape-aware Convolutional Layer for Indoor RGB-D Semantic Segmentation
    :star:code
  • Generalize then Adapt: Source-Free Domain Adaptive Semantic Segmentation
    :house:project
  • Mining Contextual Information Beyond Image for Semantic Segmentation
    :star:code
  • ISNet: Integrate Image-Level and Semantic-Level Context for Semantic Segmentation
    :star:code
  • Pseudo-mask Matters in Weakly-supervised Semantic Segmentation
    :star:code
  • SIGN: Spatial-information Incorporated Generative Network for Generalized Zero-shot Semantic Segmentation
  • Region-Aware Contrastive Learning for Semantic Segmentation
  • GP-S3Net: Graph-Based Panoptic Sparse Semantic Segmentation Network
  • Domain Adaptive Semantic Segmentation With Self-Supervised Depth Estimation
    :star:code
  • Scribble-Supervised Semantic Segmentation by Uncertainty Reduction on Neural Representation and Self-Supervision on Neural Eigenspace
  • Exploring Cross-Image Pixel Contrast for Semantic Segmentation
    :open_mouth:oral:star:code
  • Dynamic Divide-and-Conquer Adversarial Training for Robust Semantic Segmentation
    :star:code
  • Uncertainty-Aware Pseudo Label Refinery for Domain Adaptive Semantic Segmentation
  • Contrastive Learning for Label Efficient Semantic Segmentation
  • Scaling Semantic Segmentation Beyond 1K Classes on a Single GPU
    :star:code
  • Prototypical Matching and Open Set Rejection for Zero-Shot Semantic Segmentation
  • Geometric Unsupervised Domain Adaptation for Semantic Segmentation
  • Calibrated Adversarial Refinement for Stochastic Semantic Segmentation
    :star:code
  • Multi-View Radar Semantic Segmentation
    :star:code
  • Exploring Robustness of Unsupervised Domain Adaptation in Semantic Segmentation
    :open_mouth:oral:star:code
  • Specialize and Fuse: Pyramidal Output Representation for Semantic Segmentation
  • Unsupervised Semantic Segmentation by Contrasting Object Mask Proposals
    :star:code
  • Scribble-Supervised Semantic Segmentation Inference
  • Semi-Supervised Semantic Segmentation With Pixel-Level Contrastive Learning From a Class-Wise Memory Bank
    :star:code
  • 小样本语义分割
    • Simpler is Better: Few-shot Semantic Segmentation with Classifier Weight Transformer
      :star:code
    • Learning Meta-class Memory for Few-Shot Semantic Segmentation
    • Few-Shot Semantic Segmentation With Cyclic Memory Network
  • 3D语义分割
    • VMNet: Voxel-Mesh Network for Geodesic-Aware 3D Semantic Segmentation
      :open_mouth:oral:star:code
    • Sparse-to-Dense Feature Matching: Intra and Inter Domain Cross-Modal Learning in Domain Adaptation for 3D Semantic Segmentation
      :star:code
    • Weakly Supervised 3D Semantic Segmentation Using Cross-Image Consensus and Inter-Voxel Affinity Relations
  • 视频语义分割
    • Domain Adaptive Video Segmentation via Temporal Consistency Regularization
      :star:code
  • 弱监督语义分割
    • Leveraging Auxiliary Tasks with Affinity Learning for Weakly Supervised Semantic Segmentation
      :star:code
    • Complementary Patch for Weakly Supervised Semantic Segmentation
    • ECS-Net: Improving Weakly Supervised Semantic Segmentation by Using Connections Between Class Activation Maps
    • Unlocking the Potential of Ordinary Classifier: Class-Specific Adversarial Erasing Framework for Weakly Supervised Semantic Segmentation
      :star:code
    • Context Decoupling Augmentation for Weakly Supervised Semantic Segmentation
      :star:code
    • Seminar Learning for Click-Level Weakly Supervised Semantic Segmentation
  • 点云语义分割
    • ReDAL: Region-based and Diversity-aware Active Learning for Point Cloud Semantic Segmentation
      :tv:video
    • Perturbed Self-Distillation: Weakly Supervised Large-Scale Point Cloud Semantic Segmentation
    • TempNet: Online Semantic Segmentation on Large-Scale Point Cloud Series
    • Guided Point Contrastive Learning for Semi-Supervised Point Cloud Semantic Segmentation
    • Learning With Noisy Labels for Robust Point Cloud Segmentation
      :star:code:house:project
  • OOD
    • Entropy Maximization and Meta Classification for Out-of-Distribution Detection in Semantic Segmentation
      :star:code
• 实例分割
  • Rank & Sort Loss for Object Detection and Instance Segmentation
    :open_mouth:oral:star:code
  • SOTR: Segmenting Objects with Transformers
    :star:code
  • A Weakly Supervised Amodal Segmenter with Boundary Uncertainty Estimation
  • Instances as Queries
    :star:code:tv:video
  • CrossVIS: Crossover Learning for Fast Online Video Instance Segmentation
    :star:code:tv:video
  • CDNet: Centripetal Direction Network for Nuclear Instance Segmentation
    :star:code
  • PrimitiveNet: Primitive Instance Segmentation With Local Primitive Embedding Under Adversarial Metric
    :star:code
  • FASA: Feature Augmentation and Sampling Adaptation for Long-Tailed Instance Segmentation
    :star:code:house:project
  • Prior to Segment: Foreground Cues for Weakly Annotated Classes in Partially Supervised Instance Segmentation
    :star:code
  • DiscoBox: Weakly Supervised Instance Segmentation and Semantic Correspondence From Box Supervision
  • End-to-End Video Instance Segmentation via Spatial-Temporal Graph Neural Networks
  • The Surprising Impact of Mask-Head Architecture on Novel Class Segmentation
    :house:project
  • How Shift Equivariance Impacts Metric Learning for Instance Segmentation
    :star:code
  • Parallel Detection-and-Segmentation Learning for Weakly Supervised Instance Segmentation
  • Real-Time Instance Segmentation With Discriminative Orientation Maps
    :star:code
  • 视频实例分割
    • Video Instance Segmentation with a Propose-Reduce Paradigm
      :star:code
  • 3D实例分割
    • Hierarchical Aggregation for 3D Instance Segmentation
      :star:code
• 小样本分割
  • Mining Latent Classes for Few-shot Segmentation
    :open_mouth:oral:star:code
• Human Motion Segmentation(人体运动分割)
  • Graph Constrained Data Representation Learning for Human Motion Segmentation
    :star:code
  • Hypercorrelation Squeeze for Few-Shot Segmenation
    :star:code:house:project
• 点云分割
  • Learning with Noisy Labels for Robust Point Cloud Segmentation
    :open_mouth:oral:star:code:house:project
  • DRINet: A Dual-Representation Iterative Learning Network for Point Cloud Segmentation
  • RPVNet: A Deep and Efficient Range-Point-Voxel Fusion Network for LiDAR Point Cloud Segmentation
• 视频目标分割(VOS)
  • Full-Duplex Strategy for Video Object Segmentation
    :house:project
  • Joint Inductive and Transductive Learning for Video Object Segmentation
    :star:code
  • Hierarchical Memory Matching Network for Video Object Segmentation
    :star:code
  • Self-supervised Video Object Segmentation by Motion Grouping
    :star:code:house:project:tv:video
  • Deep Transport Network for Unsupervised Video Object Segmentation
  • Generating Masks From Boxes by Mining Spatio-Temporal Consistencies in Videos
    :star:code
  • Learning Motion-Appearance Co-Attention for Zero-Shot Video Object Segmentation
    :star:code
  • Video Object Segmentation With Dynamic Memory Networks and Adaptive Object Alignment
    :star:code
• 语义场景分割
  • BiMaL: Bijective Maximum Likelihood Approach to Domain Adaptation in Semantic Scene Segmentation
    :star:code
• Referring Segmentation(基于文本的分割)
  • Vision-Language Transformer and Query Generation for Referring Segmentation
    :star:code
• 场景理解
  • DeepPanoContext: Panoramic 3D Scene Understanding with Holistic Scene Context Graph and Relation-based Optimization
    :open_mouth:oral:star:code:house:project:tv:video
  • ACDC: The Adverse Conditions Dataset With Correspondences for Semantic Driving Scene Understanding
    :house:project
  • Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understanding
    :star:code
• CMA
  • Towards Better Explanations of Class Activation Mapping
  • Keep CALM and Improve Visual Feature Attribution
    :star:code
  • LFI-CAM: Learning Feature Importance for Better Visual Explanation
    :star:code
• 多目标分割
  • Faster Multi-Object Segmentation Using Parallel Quadratic Pseudo-Boolean Optimization
• 动作分割
  • Weakly-Supervised Action Segmentation and Alignment via Transcript-Aware Union-of-Subspaces Learning
    :star:code
  • Refining Action Segmentation with Hierarchical Video Representations
    :star:code
• 场景解析
  • Interaction via Bi-Directional Graph of Semantic Region Affinity for Scene Parsing
• 抠图
  • Cascade Image Matting With Deformable Graph Refinement
  • Tripartite Information Mining and Integration for Image Matting
• 运动分割
  • SLIM: Self-Supervised LiDAR Scene Flow and Motion Segmentation

12.Image/Fine-Grained Classification(图像/细粒度分类)

• DiagViB-6: A Diagnostic Benchmark Suite for Vision Models in the Presence of Shortcut and Generalization Opportunities
• Online Continual Learning For Visual Food Classification
• A Unified Objective for Novel Class Discovery
  :open_mouth:oral:star:code:house:project
  :newspaper:解读:ICCV2021 Oral | UNO：用于“新类发现”的统一目标函数，简化训练流程！已开源！
• Improving Generalization of Batch Whitening by Convolutional Unit Optimization
  :star:code
• Towards Learning Spatially Discriminative Feature Representations
• CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification
  :star:code
  :newspaper:解读:ICCV2021 MIT-IBM沃森开源CrossViT：Transformer走向多分支、多尺度
• SCOUTER: Slot Attention-based Classifier for Explainable Image Recognition
  :star:code
• Influence-Balanced Loss for Imbalanced Visual Classification
  :star:code
• Explanations for Occluded Images
  :star:code:house:project:tv:video
• Understanding Robustness of Transformers for Image Classification
• Learning Rare Category Classifiers on a Tight Labeling Budget
• Discover the Unknown Biased Attribute of an Image Classifier
  :star:code
• Co-Scale Conv-Attentional Image Transformers
  :open_mouth:oral:star:code
• Benchmark Platform for Ultra-Fine-Grained Visual Categorization Beyond Human Performance
  :star:code
• Do Image Classifiers Generalize Across Time?
  :house:project
• Interpretable Image Recognition by Constructing Transparent Embedding Space
  :star:code
• The Pursuit of Knowledge: Discovering and Localizing Novel Categories using Dual Memory
• 长尾识别
  • Parametric Contrastive Learning
    :star:code
  • ACE: Ally Complementary Experts for Solving Long-Tailed Recognition in One-Shot
    :open_mouth:oral:star:code
  • Self Supervision to Distillation for Long-Tailed Visual Recognition
    :star:code
  • Distilling Virtual Examples for Long-Tailed Recognition
  • Distributional Robustness Loss for Long-Tail Learning
  • GistNet: A Geometric Structure Transfer Network for Long-Tailed Recognition
  • 长尾视觉关系识别
    • Exploring Long Tail Visual Relationship Recognition With Large Vocabulary
      :star:code
• 细粒度
  • Webly Supervised Fine-Grained Recognition: Benchmark Datasets and An Approach
    :star:code
  • Learning Canonical 3D Object Representation for Fine-Grained Recognition
  • Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification
    :star:code
  • N-ImageNet: Towards Robust, Fine-Grained Object Recognition With Event Cameras
  • Grafit: Learning fine-grained image representations with coarse labels
  • Stochastic Partial Swap: Enhanced Model Generalization and Interpretability for Fine-Grained Recognition
    :star:code
• 小样本分类
  • Transductive Few-Shot Classification on the Oblique Manifold
  • Relational Embedding for Few-Shot Classification
    :star:code:house:project
  • Binocular Mutual Learning for Improving Few-shot Classification
    :star:code
  • Partner-Assisted Learning for Few-Shot Image Classification
  • On the Importance of Distractors for Few-Shot Classification
    :star:code
  • Few-Shot Image Classification: Just Use a Library of Pre-Trained Feature Extractors and a Simple Classifier
  • Universal Representation Learning From Multiple Domains for Few-Shot Classification
    :star:code
  • A Multi-Mode Modulator for Multi-Domain Few-Shot Classification
  • Variational Feature Disentangling for Fine-Grained Few-Shot Classification
    :star:code
  • Mixture-Based Feature Space Learning for Few-Shot Image Classification
    :star:code:house:project:tv:video
• 多标签分类
  • Asymmetric Loss For Multi-Label Classification
    :star:code
  • Semantic Diversity Learning for Zero-Shot Multi-label Classification
    :star:code

11.Visual Question Answering(视觉问答)

• Greedy Gradient Ensemble for Robust Visual Question Answering
  :star:code
• Weakly Supervised Relative Spatial Reasoning for Visual Question Answering
  :star:code
• Calibrating Concepts and Operations: Towards Symbolic Reasoning on Real Images
  :star:code
• Unshuffling Data for Improved Generalization in Visual Question Answering
• TRAR: Routing the Attention Spans in Transformer for Visual Question Answering(https://github.com/rentainhe/TRAR-VQA/)
• Contrast and Classify: Training Robust VQA Models
• Linguistically Routing Capsule Network for Out-of-Distribution Visual Question Answering
• Beyond Question-Based Biases: Assessing Multimodal Shortcut Learning in Visual Question Answering
  :star:code
• Auto-Parsing Network for Image Captioning and Visual Question Answering
• video question answering
  • Just Ask: Learning to Answer Questions from Millions of Narrated Videos
    :open_mouth:oral:star:code:house:project
  • Adversarial VQA: A New Benchmark for Evaluating the Robustness of VQA Models
    :house:project
  • Env-QA: A Video Question Answering Benchmark for Comprehensive Understanding of Dynamic Environments
    :sunflower:dataset
  • On the Hidden Treasure of Dialog in Video Question Answering
    :star:code:house:project
  • HAIR: Hierarchical Visual-Semantic Relational Reasoning for Video Question Answering
  • Video Question Answering Using Language-Guided Deep Compressed-Domain Video Feature
    :star:code
• A-VQA
  • Pano-AVQA: Grounded Audio-Visual Question Answering on 360∘ Videos

10.OCR

• Joint Visual Semantic Reasoning: Multi-Stage Decoder for Text Recognition
  :tv:video
• Text is Text, No Matter What: Unifying Text Recognition using Knowledge Distillation
  :tv:video
• Towards the Unseen: Iterative Text Recognition by Distilling from Errors
  :tv:video
• 任意形状文本检测
  • Adaptive Boundary Proposal Network for Arbitrary Shape Text Detection
    :star:code
• 场景文本识别
  • From Two to One: A New Scene Text Recognizer with Visual Language Modeling Network
    :star:code
• 场景文本替换
  • STRIVE: Scene Text Replacement In Videos
    :house:project
• 提取文档图像
  • End-to-End Piece-Wise Unwarping of Document Images
    :house:project
• 手写文本生成
  • Handwriting Transformers
    :star:code
• Table Structure Recognition(表格结构识别)
  • TGRNet: A Table Graph Reconstruction Network for Table Structure Recognition
    :star:code

9.Video

• Action Detection and Recognition(人体动作检测与识别)
  • Channel-wise Topology Refinement Graph Convolution for Skeleton-Based Action Recognition
    :star:code
  • MGSampler: An Explainable Sampling Strategy for Video Action Recognition
    :star:code
  • Skeleton Cloud Colorization for Unsupervised 3D Action Representation Learning
  • Video Pose Distillation for Few-Shot, Fine-Grained Sports Action Recognition
  • Class Semantics-based Attention for Action Detection
  • MultiSports: A Multi-Person Video Dataset of Spatio-Temporally Localized Sports Actions
    :star:code
  • AdaSGN: Adapting Joint Number and Model Size for Efficient Skeleton-Based Action Recognition
    :star:code
  • OadTR: Online Action Detection With Transformers
    :star:code
  • Self-Supervised 3D Skeleton Action Representation Learning With Motion Consistency and Continuity
  • Interactive Prototype Learning for Egocentric Action Recognition
  • Efficient Action Recognition via Dynamic Knowledge Propagation
  • Else-Net: Elastic Semantic Network for Continual Action Recognition From Skeleton Data
  • Learning Self-Similarity in Space and Time As Generalized Motion for Video Action Recognition
    :star:code:house:project
  • Temporal Action Detection With Multi-Level Supervision
    :star:code
  • Watch Only Once: An End-to-End Video Action Detection Framework
    :star:code
  • Unsupervised Few-Shot Action Recognition via Action-Appearance Aligned Meta-Adaptation
    :open_mouth:oral
  • Geometric Deep Neural Network Using Rigid and Non-Rigid Transformations for Human Action Recognition
  • Just One Moment: Structural Vulnerability of Deep Action Recognition Against One Frame Attack
  • Evidential Deep Learning for Open Set Action Recognition
    :star:code:house:project:tv:video
  • Learning an Augmented RGB Representation With Cross-Modal Knowledge Distillation for Action Detection
  • Class-Incremental Learning for Action Recognition in Videos
  • D2-Net: Weakly-Supervised Action Localization via Discriminative Embeddings and Denoised Activations
    :star:code
  • 零样本动作识别
    • Elaborative Rehearsal for Zero-shot Action Recognition
      :star:code
  • Temporal Action Localization(时序动作定位)
    • Enriching Local and Global Contexts for Temporal Action Localization
      :star:code
    • Learning Action Completeness from Points for Weakly-supervised Temporal Action Localization
      :open_mouth:oral:star:code
    • Foreground-Action Consistency Network for Weakly Supervised Temporal Action Localization
      :star:code
    • Video Self-Stitching Graph Network for Temporal Action Localization
    • Divide and Conquer for Single-Frame Temporal Action Localization
    • CAG-QIL: Context-Aware Actionness Grouping via Q Imitation Learning for Online Temporal Action Localization
  • Temporal Action Proposal Generation(时序动作提案生成)
    • Relaxed Transformer Decoders for Direct Action Proposal Generation
      :star:code
• Action Quality Assessment(行动质量评估)
  • Group-aware Contrastive Regression for Action Quality Assessment
• Video Rescaling
  • Self-Conditioned Probabilistic Learning of Video Rescaling
    :newspaper:解读:ICCV2021 |上交、北理、百度联合研究视频缩放任务中的自条件概率学习
• Video activity localisation
  • Cross-Sentence Temporal and Semantic Relations in Video Activity Localisation
    :newspaper:解读:ICCV2021 | 如何高效视频定位？QMUL&北大&Adobe强强联手提出弱监督CRM，性能SOTA
• 视频修复
  • Internal Video Inpainting by Implicit Long-range Propagation
    :star:code:house:project
  • Occlusion-Aware Video Object Inpainting
    :house:project
  • FuseFormer: Fusing Fine-Grained Information in Transformers for Video Inpainting
    :star:code
  • Flow-Guided Video Inpainting with Scene Templates
    :star:code
  • Frequency-Aware Spatiotemporal Transformers for Video Inpainting Detection
• 视频分析
  • 视频表征学习
    • Enhancing Self-supervised Video Representation Learning via Multi-level Feature Optimization
      :star:code
    • Self-Supervised Video Representation Learning with Meta-Contrastive Network
    • Long Short View Feature Decomposition via Contrastive Video Representation Learning
• 视频剪辑
  • Learning to Cut by Watching Movies
    :star:code:house:project
• 视频字幕
  • Motion Guided Region Message Passing for Video Captioning
  • Aligning Subtitles in Sign Language Videos
    :house:project:tv:video
  • Dense Video Captioning
    • End-to-End Dense Video Captioning with Parallel Decoding
      :star:code
      :newspaper:解读:港大&南科大提出端到端PDVC，用DETR的方法做Dense Video Captioning！简化训练流程
• 视频编码
  • Overfitting the Data: Compact Neural Video Delivery via Content-aware Feature Modulation
    :star:code
    :newspaper:解读:ICCV2021—工业界中的神经网络视频传输超分算法
• 视频生成
  • Click to Move: Controlling Video Generation with Sparse Motion
    :star:code
• Video Relation Detection(视频关系检测)
  • Social Fabric: Tubelet Compositions for Video Relation Detection
    :star:code
• Video Grounding
  • Support-Set Based Cross-Supervision for Video Grounding
• 视频精彩片段检测
  • Cross-category Video Highlight Detection via Set-based Learning
    :star:code
  • PR-Net: Preference Reasoning for Personalized Video Highlight Detection
  • HighlightMe: Detecting Highlights from Human-Centric Videos
  • Temporal Cue Guided Video Highlight Detection With Low-Rank Audio-Visual Fusion
  • Joint Visual and Audio Learning for Video Highlight Detection
• 视频识别
  • Searching for Two-Stream Models in Multivariate Space for Video Recognition
  • Adaptive Focus for Efficient Video Recognition
    :open_mouth:oral:star:code
  • AdaMML: Adaptive Multi-Modal Learning for Efficient Video Recognition
    :star:code:house:project
  • TAM: Temporal Adaptive Module for Video Recognition
    :star:code
  • Condensing a Sequence to One Informative Frame for Video Recognition
  • VideoLT: Large-Scale Long-Tailed Video Recognition
    :star:code
  • Motion-Augmented Self-Training for Video Recognition at Smaller Scale
  • Multi-Modal Multi-Action Video Recognition
    :star:code
• Motion Retargeting(运动重定位)
  • Contact-Aware Retargeting of Skinned Motion
    :tv:video
• 视频预测
  • A Hierarchical Variational Neural Uncertainty Model for Stochastic Video Prediction
    :open_mouth:oral
• 视频合成
  • iPOKE: Poking a Still Image for Controlled Stochastic Video Synthesis
    :star:code:house:project
• 视频帧插值
  • Training Weakly Supervised Video Frame Interpolation With Events
    :star:code
  • Asymmetric Bilateral Motion Estimation for Video Frame Interpolation
    :star:code
  • XVFI: eXtreme Video Frame Interpolation
    :open_mouth:oral:star:code:tv:video
• Deepfake 视频检测
  • ID-Reveal: Identity-aware DeepFake Video Detection
    :star:code
• 视频稳定
  • Hybrid Neural Fusion for Full-Frame Video Stabilization
    :star:code:house:project:tv:video
• Video Frame-level Similarity(视频帧级相似度学习)
  • Rethinking Self-supervised Correspondence Learning: A Video Frame-level Similarity Perspective
    :open_mouth:oral:star:code:house:project:tv:video
• 视频压缩
  • Online-Trained Upsampler for Deep Low Complexity Video Compression
• 视频时刻检索
  • Fast Video Moment Retrieval
• 视频摘要
  • Multiple Pairwise Ranking Networks for Personalized Video Summarization
• 视频质量评估
  • Unsupervised Curriculum Domain Adaptation for No-Reference Video Quality Assessment
    :star:code
• Video Grounding
  • STVGBert: A Visual-Linguistic Transformer Based Framework for Spatio-Temporal Video Grounding
• 视频定位
  • Zero-Shot Natural Language Video Localization
    :open_mouth:oral:star:code
• 视频推理
  • Real-Time Video Inference on Edge Devices via Adaptive Model Streaming
    :star:code
• 视频相关
  • Anonymizing Egocentric Videos
• 视频异常检测
  • Dance With Self-Attention: A New Look of Conditional Random Fields on Anomaly Detection in Videos
  • A Hybrid Video Anomaly Detection Framework via Memory-Augmented Flow Reconstruction and Flow-Guided Frame Prediction
    :star:code
    :newspaper:解读:ICCV 2021 oral 重构+预测，双管齐下提升视频异常检测性能
  • Weakly-Supervised Video Anomaly Detection With Robust Temporal Feature Magnitude Learning
    :star:code
• 视频去噪
  • Unsupervised Deep Video Denoising
    :star:code:house:project
• Video Portrait Relighting(人像视频重照明)
  • Neural Video Portrait Relighting in Real-time via Consistency Modeling
• 视频时序定位
  • Boundary-Sensitive Pre-Training for Temporal Localization in Videos
    :star:code:house:project
• 视频关联性
  • Explainable Video Entailment With Grounded Visual Evidence
• 视频抠图
  • Video Matting via Consistency-Regularized Graph Neural Networks
    :star:code
• 视频编码
  • ELF-VC: Efficient Learned Flexible-Rate Video Coding
• 识别视频中互动关系
  • Motion Guided Attention Fusion To Recognize Interactions From Videos
• 视频去模糊
  • Multi-Scale Separable Network for Ultra-High-Definition Video Deblurring
• 视频理解
  • Unified Graph Structured Models for Video Understanding
• 视频重建
  • HDR Video Reconstruction: A Coarse-to-fine Network and A Real-world Benchmark Dataset
    :star:code:house:project:tv:video:sunflower:dataset

8.Human Pose Estimation(人体姿态估计)

• Human Pose Regression with Residual Log-likelihood Estimation
  :open_mouth:oral:star:code
• Online Knowledge Distillation for Efficient Pose Estimation
• DECA: Deep viewpoint-Equivariant human pose estimation using Capsule Autoencoders
  :open_mouth:oral:star:code
• Estimating and Exploiting the Aleatoric Uncertainty in Surface Normal Estimation
  :open_mouth:oral:star:code
• Dynamical Pose Estimation
  :star:code:tv:video
• Multi-Instance Pose Networks: Rethinking Top-Down Pose Estimation
  :star:code:house:project
• Egocentric Pose Estimation From Human Vision Span
• Learning Privacy-Preserving Optics for Human Pose Estimation
  :open_mouth:oral:star:code:house:project:tv:video
• TokenPose: Learning Keypoint Tokens for Human Pose Estimation
  :star:code
• Motion Adaptive Pose Estimation from Compressed Videos
• 3D 人体姿态估计
  • PyMAF: 3D Human Pose and Shape Regression with Pyramidal Mesh Alignment Feedback Loop
    :open_mouth:oral:star:code:house:project
  • HuMoR: 3D Human Motion Model for Robust Pose Estimation
    :open_mouth:oral:house:project:tv:video
  • Probabilistic Monocular 3D Human Pose Estimation with Normalizing Flows
    :star:code:tv:video
  • Learning Skeletal Graph Neural Networks for Hard 3D Pose Estimation
    :star:code
  • EventHPE: Event-based 3D Human Pose and Shape Estimation
    :star:code
  • imGHUM: Implicit Generative Models of 3D Human Shape and Articulated Pose
  • Graph-Based 3D Multi-Person Pose Estimation Using Multi-View Images
  • Unsupervised 3D Pose Estimation for Hierarchical Dance Video Recognition
  • Learning to Regress Bodies from Images using Differentiable Semantic Rendering
    :house:project
  • Hierarchical Kinematic Probability Distributions for 3D Human Shape and Pose Estimation from Images in the Wild
    :star:code
  • 3D Human Pose Estimation With Spatial and Temporal Transformers
    :star:code:tv:video
  • PARE: Part Attention Regressor for 3D Human Body Estimation
    :star:code:house:project:tv:video
  • Learning Causal Representation for Training Cross-Domain Pose Estimator via Generative Interventions
  • UltraPose: Synthesizing Dense Pose With 1 Billion Points by Human-Body Decoupling 3D Model
    :star:code
  • Modulated Graph Convolutional Network for 3D Human Pose Estimation
    :star:code
  • Revitalizing Optimization for 3D Human Pose and Shape Estimation: A Sparse Constrained Formulation
    :star:code:house:project:tv:video
  • Estimating Egocentric 3D Human Pose in Global Space
    :house:project:tv:video
  • Camera Distortion-Aware 3D Human Pose Estimation in Video With Optimization-Based Meta-Learning
    :star:code
  • EM-POSE: 3D Human Pose Estimation From Sparse Electromagnetic Trackers
    :star:code:house:project:tv:video
  • Towards Alleviating the Modeling Ambiguity of Unsupervised Monocular 3D Human Pose Estimation
    :house:project
• SPEC: Seeing People in the Wild with an Estimated Camera
  :star:code:house:project:tv:video
• Encoder-Decoder With Multi-Level Attention for 3D Human Shape and Pose Estimation
  :star:code
• 3D姿势迁移
  • Unsupervised Geodesic-preserved Generative Adversarial Networks for Unconstrained 3D Pose Transfer
    :star:code
• 手部姿势
  • 手势合成
    • Speech Drives Templates: Co-Speech Gesture Synthesis with Learned Templates
      :star:code
  • 手势识别
    • Hand Image Understanding via Deep Multi-Task Learning
      :star:code
    • SemiHand: Semi-Supervised Hand Pose Estimation With Consistency
  • 3D 手部姿态
    • HandFoldingNet: A 3D Hand Pose Estimation Network Using Multiscale-Feature Guided Folding of a 2D Hand Skeleton
      :star:code
    • EventHands: Real-Time Neural 3D Hand Pose Estimation From an Event Stream
      :star:code:house:project:tv:video
    • Self-Supervised 3D Hand Pose Estimation from monocular RGB via Contrastive Learning
      :star:code:house:project:tv:video
  • 手部交互姿势估计
    • End-to-End Detection and Pose Estimation of Two Interacting Hands
  • 3D手网格建模
    • I2UV-HandNet: Image-to-UV Prediction Network for Accurate and High-Fidelity 3D Hand Mesh Modeling
  • Towards Accurate Alignment in Real-Time 3D Hand-Mesh Reconstruction
  • 手部网格恢复
    • Self-Supervised Transfer Learning for Hand Mesh Recovery From Binocular Images
  • 手势学习
    • TravelNet: Self-Supervised Physically Plausible Hand Motion Learning From Monocular Color Images
      :house:project
  • 手势重建
    • Interacting Two-Hand 3D Pose and Shape Reconstruction from Single Color Image
      :star:code
• 三维网格合成
  • Deep Hybrid Self-Prior for Full 3D Mesh Generation
    :house:project
  • Mesh Graphormer
    :star:code
• 人体重建
  • ARCH++: Animation-Ready Clothed Human Reconstruction Revisited
    :tv:video
  • 3D 人体重建
    • Probabilistic Modeling for Human Mesh Recovery
      :star:code:house:project
    • Gravity-Aware Monocular 3D Human-Object Reconstruction
      :house:project
    • THUNDR: Transformer-Based 3D Human Reconstruction With Markers
    • NPMs: Neural Parametric Models for 3D Deformable Shapes
      :star:code:house:project:tv:video
• 4D人体捕捉
  • Learning Motion Priors for 4D Human Body Capture in 3D Scenes
    :star:code:house:project:tv:video
• 人体姿态估计与合成
  • Physics-based Human Motion Estimation and Synthesis from Videos
• 多人姿态估计
  • Shape-aware Multi-Person Pose Estimation from Multi-View Images
    :star:code:house:project:tv:video
    论文公开
  • The Center of Attention: Center-Keypoint Grouping via Attention for Multi-Person Pose Estimation
    :star:code
• 人/物体姿态关键点检测
  • Keypoint Communities
    :star:code
• 人体运动捕捉
  • SOMA: Solving Optical Marker-Based MoCap Automatically
    :star:code:house:project:tv:video
  • DeepMultiCap: Performance Capture of Multiple Characters Using Sparse Multiview Cameras
    :house:project
  • Lightweight Multi-person Total Motion Capture Using Sparse Multi-view Cameras
• 2D人体姿势估计
  • An Empirical Study of the Collapsing Problem in Semi-Supervised 2D Human Pose Estimation
    :star:code
• Human Action Video Alignment
  • Normalized Human Pose Features for Human Action Video Alignment
• 3D姿态迁移
  • Intrinsic-Extrinsic Preserved GANs for Unsupervised 3D Pose Transfer
    :star:code
• 人体网格恢复
  • Skeleton2Mesh: Kinematics Prior Injected Unsupervised Human Mesh Recovery
    :house:project
  • Uncertainty-Aware Human Mesh Recovery From Video by Learning Part-Based 3D Dynamics
• 根据人体姿势估计距离
  • Single View Physical Distance Estimation using Human Pose
    :house:project
• 3D人体
  • SNARF: Differentiable Forward Skinning for Animating Non-Rigid Neural Implicit Shapes
    :star:code:house:project:tv:video
  • Monocular, One-stage, Regression of Multiple 3D People
    :star:code
• 运动合成
  • Synthesis of Compositional Animations from Textual Descriptions
    :star:code
  • 3D人体运动合成
    • Action-Conditioned 3D Human Motion Synthesis With Transformer VAE
    • A Unified 3D Human Motion Synthesis Model via Conditional Variational Auto-Encoder
• 3D动画
  • DeePSD: Automatic Deep Skinning and Pose Space Deformation for 3D Garment Animation
• 服装类别级姿势估计
  • GarmentNets: Category-Level Pose Estimation for Garments via Canonical Space Shape Completion
    :star:code:house:project
• 服装人体建模
  • Point-Based Modeling of Human Clothing
    :star:code:house:project:tv:video
• 关键点定位
  • TransPose: Keypoint Localization via Transformer
    :star:code

7.Scene Graph Generation(场景图生成)

• Spatial-Temporal Transformer for Dynamic Scene Graph Generation
  :star:code:tv:video
• Unconditional Scene Graph Generation
  :house:project
• Target Adaptive Context Aggregation for Video Scene Graph Generation
  :star:code
• Learning to Generate Scene Graph from Natural Language Supervision
  :star:code
• Segmentation-Grounded Scene Graph Generation
  :star:code
• Context-aware Scene Graph Generation with Seq2Seq Transformer
  :star:code
• A Simple Baseline for Weakly-Supervised Scene Graph Generation
  :star:code
• Generative Compositional Augmentations for Scene Graph Prediction
  :star:code
• From General to Specific: Informative Scene Graph Generation via Balance Adjustment
  :star:code
• 场景合成
  • Self-Supervised Real-to-Sim Scene Generation
    :house:project
  • Unconstrained Scene Generation With Locally Conditioned Radiance Fields
    :star:code:house:project
  • Visual Distant Supervision for Scene Graph Generation
    :star:code

6.Point Cloud(点云)

• AdaFit: Rethinking Learning-based Normal Estimation on Point Clouds
  :star:code:house:project
• Adaptive Graph Convolution for Point Cloud Analysis
  :star:code
• Learning Inner-Group Relations on Point Clouds
• CPFN: Cascaded Primitive Fitting Networks for High-Resolution Point Clouds
  :star:code
• Cloud Transformers: A Universal Approach To Point Cloud Processing Tasks
  :star:code:tv:video
• PCAM: Product of Cross-Attention Matrices for Rigid Registration of Point Clouds
  :star:code
• 3DVG-Transformer: Relation Modeling for Visual Grounding on Point Clouds
• Differentiable Convolution Search for Point Cloud Processing
• Superpoint Network for Point Cloud Oversegmentation
  :star:code
• PU-EVA: An Edge-Vector Based Approximation Solution for Flexible-Scale Point Cloud Upsampling
• SGMNet: Learning Rotation-Invariant Point Cloud Representations via Sorted Gram Matrix
• DWKS: A Local Descriptor of Deformations Between Meshes and Point Clouds
  :star:code
• Robustness Certification for Point Cloud Models
  :star:code
• Vector Neurons: A General Framework for SO(3)-Equivariant Networks
  :star:code
• Unsupervised Point Cloud Pre-Training via Occlusion Completion
  :star:code
• Towards Efficient Graph Convolutional Networks for Point Cloud Handling
  :star:code
• Progressive Seed Generation Auto-Encoder for Unsupervised Point Cloud Learning
• 点云去噪
  • Score-Based Point Cloud Denoising
    :star:code
• 点云配准
  • HRegNet: A Hierarchical Network for Large-scale Outdoor LiDAR Point Cloud Registration
    :star:code:house:project
  • (Just) A Spoonful of Refinements Helps the Registration Error Go Down
    :open_mouth:oral:star:code
  • A Robust Loss for Point Cloud Registration
  • Deep Hough Voting for Robust Global Registration
  • Sampling Network Guided Cross-Entropy Method for Unsupervised Point Cloud Registration
    :star:code
  • Feature Interactive Representation for Point Cloud Registration
  • LSG-CPD: Coherent Point Drift With Local Surface Geometry for Point Cloud Registration
    :star:code:tv:video
  • OMNet: Learning Overlapping Mask for Partial-to-Partial Point Cloud Registration
    :star:code
  • DeepPRO: Deep Partial Point Cloud Registration of Objects
  • Provably Approximated Point Cloud Registration
  • Bootstrap Your Own Correspondences点云配准
  • Distinctiveness Oriented Positional Equilibrium for Point Cloud Registration
• 3D点云
  • Unsupervised Learning of Fine Structure Generation for 3D Point Clouds by 2D Projection Matching
    :star:code
  • Spatio-temporal Self-Supervised Representation Learning for 3D Point Clouds
    :star:code:house:project
  • Unsupervised Learning of Fine Structure Generation for 3D Point Clouds by 2D Projections Matching
    :star:code
  • Point Transformer
  • Point-Set Distances for Learning Representations of 3D Point Clouds
  • PointBA: Towards Backdoor Attacks in 3D Point Cloud
  • Minimal Adversarial Examples for Deep Learning on 3D Point Clouds
  • 3D点云重建
    • MonteFloor: Extending MCTS for Reconstructing Accurate Large-Scale Floor Plans
      :open_mouth:oral:house:project:tv:video
• 点云补全
  • SnowflakeNet: Point Cloud Completion by Snowflake Point Deconvolution with Skip-Transformer
    :open_mouth:oral:star:code
  • ME-PCN: Point Completion Conditioned on Mask Emptiness
  • PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers
    :open_mouth:oral:star:code
  • Voxel-based Network for Shape Completion by Leveraging Edge Generation
    :star:code
  • RFNet: Recurrent Forward Network for Dense Point Cloud Completion
• 点云增强
  • Point Cloud Augmentation with Weighted Local Transformations
    :star:code
• 点云形状分析
  • Walk in the Cloud: Learning Curves for Point Clouds Shape Analysis
    :star:code:house:project
• 点云分析
  • A Closer Look at Rotation-Invariant Deep Point Cloud Analysis
    :star:code
• 3D点云分类
  • A Backdoor Attack Against 3D Point Cloud Classifiers
    :star:code
• 3D点云生成与补全
  • 3D Shape Generation and Completion through Point-Voxel Diffusion
    :house:project:tv:video
• point cloud object co-segmentation
  • Unsupervised Point Cloud Object Co-Segmentation by Co-Contrastive Learning and Mutual Attention Sampling
    :star:code
• 点云理解
  • Shape Self-Correction for Unsupervised Point Cloud Understanding

5.Few-Shot/Zero-Shot Learning;Domain Generalization/Adaptation(小/零样本学习;域适应/泛化)

• 域适应
  • Transporting Causal Mechanisms for Unsupervised Domain Adaptation
    :open_mouth:oral
    :star:code
  • Generalized Source-free Domain Adaptation
    :star:code
  • Semantic Concentration for Domain Adaptation
    :star:code
  • PIT: Position-Invariant Transform for Cross-FoV Domain Adaptation
    :star:code
  • Learning Cross-modal Contrastive Features for Video Domain Adaptation
  • Zero-Shot Day-Night Domain Adaptation With a Physics Prior
    :open_mouth:oral:star:code
  • Active Universal Domain Adaptation
  • Re-energizing Domain Discriminator with Sample Relabeling for Adversarial Domain Adaptation
  • OVANet: One-vs-All Network for Universal Domain Adaptation
    :star:code
  • Collaborative Optimization and Aggregation for Decentralized Domain Generalization and Adaptation
  • Partial Video Domain Adaptation with Partial Adversarial Temporal Attentive Network
    :star:code
  • Information-Theoretic Regularization for Multi-Source Domain Adaptation
  • Gradient Distribution Alignment Certificates Better Adversarial Domain Adaptation
  • Adaptive Adversarial Network for Source-Free Domain Adaptation
    :star:code
  • T-SVDNet: Exploring High-Order Prototypical Correlations for Multi-Source Domain Adaptation
    :star:code
  • Self-Supervised Domain Adaptation for Forgery Localization of JPEG Compressed Images
  • ECACL: A Holistic Framework for Semi-Supervised Domain Adaptation
    :star:code
  • STEM: An approach to Multi-source Domain Adaptation with Guarantees
  • Towards Novel Target Discovery Through Open-Set Domain Adaptation
    :star:code
  • Deep Co-Training With Task Decomposition for Semi-Supervised Domain Adaptation
    :star:code
  • mDALU: Multi-Source Domain Adaptation and Label Unification with Partial Datasets
  • Geometry-Aware Self-Training for Unsupervised Domain Adaptation on Object Point Clouds
    :star:code
  • 无监督域适应
    • Adversarial Unsupervised Domain Adaptation with Conditional and Label Shift: Infer, Align and Iterate
    • Recursively Conditional Gaussian for Ordinal Unsupervised Domain Adaptation
      :open_mouth:oral
    • Tune it the Right Way: Unsupervised Validation of Domain Adaptation via Soft Neighborhood Density
      :star:code
    • Adversarial Robustness for Unsupervised Domain Adaptation
      :house:project
    • SENTRY: Selective Entropy Optimization via Committee Consistency for Unsupervised Domain Adaptation
      :star:code
  • 零样本域适应
    • Zero-Shot Domain Adaptation with a Physics Prior
      :open_mouth:oral:star:code
    • Collaborative Learning with Disentangled Features for Zero-shot Domain Adaptation
• 域泛化
  • Domain Generalization via Gradient Surgery
    :star:code
  • Learning to Diversify for Single Domain Generalization
    :star:code
  • Shape-Biased Domain Generalization via Shock Graph Embeddings
  • SelfReg: Self-Supervised Contrastive Regularization for Domain Generalization
  • A Style and Semantic Memory Mechanism for Domain Generalization
  • Confidence Calibration for Domain Generalization Under Covariate Shift
  • A Simple Feature Augmentation for Domain Generalization
• 小样本
  • Boosting the Generalization Capability in Cross-Domain Few-shot Learning via Noise-enhanced Supervised Autoencoder
  • Generalized and Incremental Few-Shot Learning by Explicit Learning and Calibration without Forgetting
    :star:code
  • Meta Navigator: Search for a Good Adaptation Policy for Few-shot Learning
  • Meta-Learning with Task-Adaptive Loss Function for Few-Shot Learning
    :open_mouth:oral:star:code
  • Z-Score Normalization, Hubness, and Few-Shot Learning
  • Pseudo-Loss Confidence Metric for Semi-Supervised Few-Shot Learning
  • Curvature Generation in Curved Spaces for Few-Shot Learning
    :star:code
  • Task-Aware Part Mining Network for Few-Shot Learning
  • Meta-Baseline: Exploring Simple Meta-Learning for Few-Shot Learning
    :star:code
  • UVStyle-Net: Unsupervised Few-Shot Learning of 3D Style Similarity Measure for B-Reps
    :star:code
  • Shallow Bayesian Meta Learning for Real-World Few-Shot Recognition
    :star:code
  • Iterative Label Cleaning for Transductive and Semi-Supervised Few-Shot Learning
    :star:code
  • Coarsely-labeled Data for Better Few-shot Transfer
    :star:code
  • 小样本异常检测
    • A Hierarchical Transformation-Discriminating Generative Model for Few Shot Anomaly Detection
• Zero-Shot Learning(零样本学习)
  • Discriminative Region-based Multi-Label Zero-Shot Learning
    :star:code
  • Field-Guide-Inspired Zero-Shot Learning
  • Generalized Zero-Shot Learning(广义零样本学习)
    • FREE: Feature Refinement for Generalized Zero-Shot Learning
      :star:code
    • Semantics Disentangling for Generalized Zero-Shot Learning
      :star:code

4.Neural rendering(神经渲染)

• In-Place Scene Labelling and Understanding with Implicit Scene Representation
  :open_mouth:oral:house:project:tv:video
• Differentiable Surface Rendering via Non-Differentiable Sampling
• Self-Calibrating Neural Radiance Fields
  :star:code
• NerfingMVS: Guided Optimization of Neural Radiance Fields for Indoor Multi-view Stereo
  :open_mouth:oral:star:code:house:project
• Learning Object-Compositional Neural Radiance Field for Editable Scene Rendering
  :star:code:house:project
• CodeNeRF: Disentangled Neural Radiance Fields for Object Categories
  :star:code
• MVSNeRF: Fast Generalizable Radiance Field Reconstruction from Multi-View Stereo
  :star:code:house:project:tv:video
• PlenOctrees for Real-Time Rendering of Neural Radiance Fields
  :open_mouth:oral:star:Conversion Code:star:Viewer Code:house:project:tv:video
• Neural Radiance Flow for 4D View Synthesis and Video Processing
  :star:code:house:project
• Animatable Neural Radiance Fields for Modeling Dynamic Human Bodies
  :star:code:house:project:tv:video
  :newspaper:解读:浙大三维视觉团队提出Animatable NeRF，从RGB视频中重建可驱动人体模型 (ICCV'21)
• GNeRF: GAN-Based Neural Radiance Field Without Posed Camera
  :open_mouth:oral
• BARF: Bundle-Adjusting Neural Radiance Fields
  :open_mouth:oral:star:code:house:project
• FastNeRF: High-Fidelity Neural Rendering at 200FPS
  :house:project:tv:video
• PIRenderer: Controllable Portrait Image Generation via Semantic Neural Rendering
  :star:code:tv:video
• NeRD: Neural Reflectance Decomposition from Image Collections
  :star:code:house:project:tv:video
• Editing Conditional Radiance Fields
  :star:code:house:project:tv:video
• GRF: Learning a General Radiance Field for 3D Representation and Rendering
  :star:code
• 4DComplete: Non-Rigid Motion Estimation Beyond the Observable Surface
  :star:code:tv:video
• KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPs
  :star:code
• Neural Articulated Radiance Field
  :star:code
• Baking Neural Radiance Fields for Real-Time View Synthesis
  :house:project:tv:video
• Non-Rigid Neural Radiance Fields: Reconstruction and Novel View Synthesis of a Dynamic Scene From Monocular Video
  :star:code:house:project
• Nerfies: Deformable Neural Radiance Fields
  :star:code:house:project:tv:video
• Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields
• UNISURF: Unifying Neural Implicit Surfaces and Radiance Fields for Multi-View Reconstruction
  :open_mouth:oral:star:code:house:project:tv:video
• 3D渲染
  • GANcraft: Unsupervised 3D Neural Rendering of Minecraft Worlds
    :open_mouth:oral:star:code:house:project:tv:video
• 3D photography(3D 相片)
  • SLIDE: Single Image 3D Photography with Soft Layering and Depth-aware Inpainting
    :open_mouth:oral:house:project:tv:video
• 渲染
  • EgoRenderer: Rendering Human Avatars from Egocentric Camera Images

3.Image Clustering(图像聚类)

• Clustering by Maximizing Mutual Information Across Views
• Learning Hierarchical Graph Neural Networks for Image Clustering
  :star:code
• One-Pass Multi-View Clustering for Large-Scale Data
• End-to-End Robust Joint Unsupervised Image Alignment and Clustering
• Graph Contrastive Clustering
  :star:code
• 人脸聚类
  • Learn To Cluster Faces via Pairwise Classification

2.Sign Language(手语识别)

• Mixed SIGNals: Sign Language Production via a Mixture of Motion Primitives
• SignBERT: Pre-Training of Hand-Model-Aware Representation for Sign Language Recognition
• Self-Mutual Distillation Learning for Continuous Sign Language Recognition
• Visual Alignment Constraint for Continuous Sign Language Recognition
  :star:code
• 手语翻译
  • Stochastic Transformer Networks With Linear Competing Units: Application To End-to-End SL Translation

1.Other(其它)

• Bias Loss for Mobile Neural Networks
  :star:code
• Improve Unsupervised Pretraining for Few-label Transfer
• Temporal-wise Attention Spiking Neural Networks for Event Streams Classification
• Accelerating Atmospheric Turbulence Simulation via Learned Phase-to-Space Transform
• Energy-Based Open-World Uncertainty Modeling for Confidence Calibration
• Robustness via Cross-Domain Ensembles
  :open_mouth:oral:star:code:house:project:tv:video
• Warp Consistency for Unsupervised Learning of Dense Correspondences
  :open_mouth:oral:star:code
• Few-Shot and Continual Learning with Attentive Independent Mechanisms
  :star:code
• Out-of-Core Surface Reconstruction via Global TGV Minimization
• ELLIPSDF: Joint Object Pose and Shape Optimization with a Bi-level Ellipsoid and Signed Distance Function Description
• Multi-scale Matching Networks for Semantic Correspondence
  :star:code
• Learning with Noisy Labels via Sparse Regularization
  :star:code
• CanvasVAE: Learning to Generate Vector Graphic Documents
• Toward Spatially Unbiased Generative Models
  :star:code
• Learning Compatible Embeddings
  :star:code
• Instance Similarity Learning for Unsupervised Feature Representation
  :star:code
• Generalizable Mixed-Precision Quantization via Attribution Rank Preservation
  :star:code
• Unifying Nonlocal Blocks for Neural Networks
  :star:code
• Impact of Aliasing on Generalization in Deep Convolutional Networks
• NASOA: Towards Faster Task-oriented Online Fine-tuning with a Zoo of Models
  :star:code
• ProAI: An Efficient Embedded AI Hardware for Automotive Applications - a Benchmark Study
• m-RevNet: Deep Reversible Neural Networks with Momentum 涉嫌学术不端，已申请撤稿
• Continual Neural Mapping: Learning An Implicit Scene Representation from Sequential Observations
• MT-ORL: Multi-Task Occlusion Relationship Learning
  :star:code
• Finding Representative Interpretations on Convolutional Neural Networks
• Orthogonal Jacobian Regularization for Unsupervised Disentanglement in Image Generation
  :star:code
• PR-RRN: Pairwise-Regularized Residual-Recursive Networks for Non-rigid Structure-from-Motion
• Instance Segmentation in 3D Scenes using Semantic Superpoint Tree Networks
  :star:code
• Learning RAW-to-sRGB Mappings with Inaccurately Aligned Supervision
  :star:code
• Structured Outdoor Architecture Reconstruction by Exploration and Classification
• Global Pooling, More than Meets the Eye: Position Information is Encoded Channel-Wise in CNNs
  :star:code
• A New Journey from SDRTV to HDRTV
  :star:code
• A Simple Framework for 3D Lensless Imaging with Programmable Masks
  :star:code
• Causal Attention for Unbiased Visual Recognition
  :star:code
• Learning to Match Features with Seeded Graph Matching Network
  :star:code
• Amplitude-Phase Recombination: Rethinking Robustness of Convolutional Neural Networks in Frequency Domain
• PatchMatch-RL: Deep MVS with Pixelwise Depth, Normal, and Visibility
  :open_mouth:oral
• Towards Understanding the Generative Capability of Adversarially Robust Classifiers
  :open_mouth:oral
• Ranking Models in Unlabeled New Environments
  :star:code
• Learning of Visual Relations: The Devil is in the Tails
  :house:project
• BlockCopy: High-Resolution Video Processing with Block-Sparse Feature Propagation and Online Policies
  :star:code
• Patch2CAD: Patchwise Embedding Learning for In-the-Wild Shape Retrieval from a Single Image
• 去偏差
  • BiaSwap: Removing dataset bias with bias-tailored swapping augmentation
• Full-Velocity Radar Returns by Radar-Camera Fusion
• CSG-Stump: A Learning Friendly CSG-Like Representation for Interpretable Shape Parsing
  :star:code:house:project
• NGC: A Unified Framework for Learning with Open-World Noisy Data
• LocTex: Learning Data-Efficient Visual Representations from Localized Textual Supervision
  :house:project
• Unsupervised Dense Deformation Embedding Network for Template-Free Shape Correspondence
• Lifelong Infinite Mixture Model Based on Knowledge-Driven Dirichlet Process
  :star:code
• Digging into Uncertainty in Self-supervised Multi-view Stereo
• Learning to Discover Reflection Symmetry via Polar Matching Convolution
  :star:code:house:project
• A Dual Adversarial Calibration Framework for Automatic Fetal Brain Biometry
• The Functional Correspondence Problem
• The Animation Transformer: Visual Correspondence via Segment Matching
• Parsing Table Structures in the Wild
  :star:code
• Square Root Marginalization for Sliding-Window Bundle Adjustment
  :star:code:house:project:tv:video
• Hierarchical Object-to-Zone Graph for Object Navigation
  :star:code:tv:video
• Robustness and Generalization via Generative Adversarial Training
• Learning Fast Sample Re-weighting Without Reward Data
  :star:code
• ReconfigISP: Reconfigurable Camera Image Processing Pipeline
  :house:project
• Learning Indoor Inverse Rendering with 3D Spatially-Varying Lighting
  :open_mouth:oral
• Low-Shot Validation: Active Importance Sampling for Estimating Classifier Performance on Rare Categories
• DisUnknown: Distilling Unknown Factors for Disentanglement Learning
  :star:code:house:project
• S3VAADA: Submodular Subset Selection for Virtual Adversarial Active Domain Adaptation
  :house:project
• ALADIN: All Layer Adaptive Instance Normalization for Fine-grained Style Similarity
  :tv:video
• Photon-Starved Scene Inference using Single Photon Cameras
  :tv:video
• OSCAR-Net: Object-centric Scene Graph Attention for Image Attribution
  :star:code:house:project
• Learning to Estimate Hidden Motions with Global Motion Aggregation
  :star:code:tv:video
• Modelling Neighbor Relation in Joint Space-Time Graph for Video Correspondence Learning
• Meta Learning on a Sequence of Imbalanced Domains with Difficulty Awareness
  :star:code
• Procedure Planning in Instructional Videosvia Contextual Modeling and Model-based Policy Learning
  :open_mouth:oral
• Extensions of Karger's Algorithm: Why They Fail in Theory and How They Are Useful in Practice
  :open_mouth:oral:star:code
• Neural Strokes: Stylized Line Drawing of 3D Shapes
  :star:code
• Learning Realistic Human Reposing using Cyclic Self-Supervision with 3D Shape, Pose, and Appearance Consistency
• Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets from 3D Scans
  :house:project
• Cherry-Picking Gradients: Learning Low-Rank Embeddings of Visual Data via Differentiable Cross-Approximation
  :star:code
• Exploiting Explanations for Model Inversion Attacks
• Learning Bias-Invariant Representation by Cross-Sample Mutual Information Minimization
• RDI-Net: Relational Dynamic Inference Networks
  :star:code
• ARAPReg: An As-Rigid-As Possible Regularization Loss for Learning Deformable Shape Generators
  :star:code
• T-Net: Effective Permutation-Equivariant Network for Two-View Correspondence Learning
  :star:code
• Learning To Stylize Novel Views
  :star:code:house:project
• A Lazy Approach to Long-Horizon Gradient-Based Meta-Learning
• Viewing Graph Solvability via Cycle Consistency
  :open_mouth:oral:star:code
  :trophy:Best paper honorable mention
• SACoD: Sensor Algorithm Co-Design Towards Efficient CNN-Powered Intelligent PhlatCam
  :star:code
• Rethinking 360° Image Visual Attention Modelling with Unsupervised Learning
• Motion Basis Learning for Unsupervised Deep Homography Estimation with Subspace Projection
  :star:code
• Batch Normalization Increases Adversarial Vulnerability and Decreases Adversarial Transferability: A Non-Robust Feature Perspective
• DeepCAD: A Deep Generative Network for Computer-Aided Design Models
  :house:project
• Better Aggregation in Test-Time Augmentation
• Self-Born Wiring for Neural Trees
• Detector-Free Weakly Supervised Grounding by Separation
• Motion-Aware Dynamic Architecture for Efficient Frame Interpolation
• Relating Adversarially Robust Generalization to Flat Minima
• Bit-Mixer: Mixed-Precision Networks With Runtime Bit-Width Selection
• AINet: Association Implantation for Superpixel Segmentation
  :star:code
• Orthogonal Projection Loss
  :star:code
• Knowledge-Enriched Distributional Model Inversion Attacks
  :star:code
• Architecture Disentanglement for Deep Neural Networks
  :star:code
• On Equivariant and Invariant Learning of Object Landmark Representations
  :star:code:house:project
• Predicting with Confidence on Unseen Distributions
• Embed Me If You Can: A Geometric Perceptron
  :star:code
• Persistent Homology Based Graph Convolution Network for Fine-Grained 3D Shape Segmentation
• HIRE-SNN: Harnessing the Inherent Robustness of Energy-Efficient Deep Spiking Neural Networks by Training With Crafted Input Noise
  :star:code
• Towards Memory-Efficient Neural Networks via Multi-Level In Situ Generation
• From Culture to Clothing: Discovering the World Events Behind a Century of Fashion Images
  :house:project
• MBA-VO: Motion Blur Aware Visual Odometry
  :star:code
• STR-GQN: Scene Representation and Rendering for Unknown Cameras Based on Spatial Transformation Routing
• Explaining Local, Global, And Higher-Order Interactions In Deep Learning
• Beyond Trivial Counterfactual Explanations with Diverse Valuable Explanations
  :star:code
• Homogeneous Architecture Augmentation for Neural Predictor
  :star:code
• SS-IL: Separated Softmax for Incremental Learning
• VSAC: Efficient and Accurate Estimator for H and F
• Fusion Moves for Graph Matching
  :star:code:house:project
• Geometric Granularity Aware Pixel-To-Mesh
• Modulated Periodic Activations for Generalizable Local Functional Representations
  :house:project
• Interpretation of Emergent Communication in Heterogeneous Collaborative Embodied Agents
  :star:code:house:project:tv:video
• A Dark Flash Normal Camera
  :house:project:tv:video
• Pri3D: Can 3D Priors Help 2D Representation Learning?
  :star:code:tv:video
• Membership Inference Attacks Are Easier on Difficult Problems
• Auxiliary Tasks and Exploration Enable ObjectGoal Navigation
  :star:code:house:project
• MixMo: Mixing Multiple Inputs for Multiple Outputs via Deep Subnetworks
• Act the Part: Learning Interaction Strategies for Articulated Object Part Discovery
  :house:project
• DCT-SNN: Using DCT To Distribute Spatial Information Over Time for Low-Latency Spiking Neural Networks
  :star:code
• Learning To Resize Images for Computer Vision Tasks
• Field of Junctions: Extracting Boundary Structure at Low SNR
• DeepGaze IIE: Calibrated prediction in and out-of-domain for state-of-the-art saliency modeling
• Learning To Reduce Defocus Blur by Realistically Modeling Dual-Pixel Data
  :star:code
• Graph-based Asynchronous Event Processing for Rapid Object Recognitio
• Ranking Models in Unlabeled New Environments
  :star:code
• A Hybrid Frequency-Spatial Domain Model for Sparse Image Reconstruction in Scanning Transmission Electron Microscopy
  :star:code
• MixMix: All You Need for Data-Free Compression Are Feature and Data Mixing
• Efficient Large Scale Inlier Voting for Geometric Vision Problems
  :star:code
• Aggregation With Feature Detection
• ReCU: Reviving the Dead Weights in Binary Neural Networks
  :star:code
• Deep Halftoning With Reversible Binary Pattern
• FFT-OT: A Fast Algorithm for Optimal Transportation
• Progressive Correspondence Pruning by Consensus Learning
  :star:code:house:project
  :newspaper:解读:基于一致性学习的渐进式匹配筛选 (ICCV 2021)
• Multispectral Illumination Estimation Using Deep Unrolling Network
• Distilling Global and Local Logits With Densely Connected Relations
• Learning specialized activation functions with the Piecewise Linear Unit
• Adaptive Convolutions With Per-Pixel Dynamic Filter Atom
• Deep Matching Prior: Test-Time Optimization for Dense Correspondence
  :star:code
• Calibrated and Partially Calibrated Semi-Generalized Homographies
  :star:code
• The Spatio-Temporal Poisson Point Process: A Simple Model for the Alignment of Event Camera Data
  :star:code
• EC-DARTS: Inducing Equalized and Consistent Optimization Into DARTS
• Refining activation downsampling with SoftPool
• FATNN: Fast and Accurate Ternary Neural Networks
  :star:code
• GTT-Net: Learned Generalized Trajectory Triangulation
• Deep Permutation Equivariant Structure from Motion
  :star:code
• Extending Neural P-frame Codecs for B-frame Codin
• Hierarchical Graph Attention Network for Few-Shot Visual-Semantic Learning
• SA-ConvONet: Sign-Agnostic Optimization of Convolutional Occupancy Networks
  :star:code
• AA-RMVSNet: Adaptive Aggregation Recurrent Multi-View Stereo Network
  :star:code
• Rethinking the Backdoor Attacks' Triggers: A Frequency Perspective
  :star:code
• Orthographic-Perspective Epipolar Geometry
• Why Approximate Matrix Square Root Outperforms Accurate SVD in Global Covariance Pooling?
  :star:code
• PixelPyramids: Exact Inference Models From Lossless Image Pyramids
  :star:code
• SurfaceNet: Adversarial SVBRDF Estimation from a Single Image
  :star:code
• Adaptive Curriculum Learning
• Sparse-Shot Learning With Exclusive Cross-Entropy for Extremely Many Localisations
• Graspness Discovery in Clutters for Fast and Accurate Grasp Detection
• RobustNav: Towards Benchmarking Robustness in Embodied Navigation
  :star:code
• Generating Attribution Maps With Disentangled Masked Backpropagation
• Spectral Leakage and Rethinking the Kernel Size in CNNs
  :star:code
• What You Can Learn by Staring at a Blank Wall
• Neural TMDlayer: Modeling Instantaneous Flow of Features via SDE Generators
• CLEAR: Clean-up Sample-Targeted Backdoor in Neural Networks
• Learning To Hallucinate Examples From Extrinsic and Intrinsic Supervision
• Single-shot Hyperspectral-Depth Imaging with Learned Diffractive Optics
• GridToPix: Training Embodied Agents With Minimal Supervision
  :house:project:tv:video
• Differentiable Dynamic Wirings for Neural Networks
• JEM++: Improved Techniques for Training JEM
  :star:code
• X-World: Accessibility, Vision, and Autonomy Meet
• Memory-augmented Dynamic Neural Relational Inference
• Physics-based Differentiable Depth Sensor Simulation
• Hypergraph Neural Networks for Hypergraph Matching
  :star:code
• Visual Grounding
  • SAT: 2D Semantics Assisted Training for 3D Visual Grounding
• Cortical Surface Shape Analysis Based on Alexandrov Polyhedra
• FcaNet: Frequency Channel Attention Networks
  :star:code
• Procedure Planning in Instructional Videos via Contextual Modeling and Model-Based Policy Learning
• Structured Outdoor Architecture Reconstruction by Exploration and Classification
  :star:code
• ELLIPSDF: Joint Object Pose and Shape Optimization with a Bi-level Ellipsoid and Signed Distance Function Description
• Testing Using Privileged Information by Adapting Features With Statistical Dependence
• Virtual Light Transport Matrices for Non-Line-of-Sight Imaging
  :open_mouth:oral
• DecentLaM: Decentralized Momentum SGD for Large-batch Deep Training
• Contrastive Multimodal Fusion with TupleInfoNCE
• Learning Better Visual Data Similarities via New Grouplet Non-Euclidean Embedding
• An Elastica Geodesic Approach With Convexity Shape Prior
• Inverting a Rolling Shutter Camera: Bring Rolling Shutter Images to High Framerate Global Shutter Video
• Multimodal Knowledge Expansion
  :star:code
• Direct Differentiable Augmentation Search
  :star:code
• The Functional Correspondence Problem
  :house:project
• Joint Topology-Preserving and Feature-Refinement Network for Curvilinear Structure Segmentation
  :star:code
• Generative Layout Modeling Using Constraint Graphs
• Self-Supervised Image Prior Learning with GMM from a Single Noisy Image
  :star:code
• Deep Implicit Surface Point Prediction Networks
  :star:code:house:project:tv:video
• Poly-NL: Linear Complexity Non-local Layers With 3rd Order Polynomials
• Factorizing Perception and Policy for Interactive Instruction Following
  :star:code
• Group-Wise Inhibition Based Feature Regularization for Robust Classification
  :star:code
• Searching for Robustness: Loss Learning for Noisy Classification Tasks
• Statistically Consistent Saliency Estimation
• Practical Relative Order Attack in Deep Ranking
  :star:code
• Q-Match: Iterative Shape Matching via Quantum Annealing
  :star:code:house:project
• Learning To Better Segment Objects From Unseen Classes With Unlabeled Videos
  :house:project:tv:video
• Globally Optimal and Efficient Manhattan Frame Estimation by Delimiting Rotation Search Space
• Cross-Encoder for Unsupervised Gaze Representation Learning
• Hierarchical Disentangled Representation Learning for Outdoor Illumination Estimation and Editing
• NeuSpike-Net: High Speed Video Reconstruction via Bio-Inspired Neuromorphic Cameras
• Local Temperature Scaling for Probability Calibration
• LIRA: Learnable, Imperceptible and Robust Backdoor Attacks
• Conformer: Local Features Coupling Global Representations for Visual Recognition
  :star:code
• Reliably fast adversarial training via latent adversarial perturbation
• PX-NET: Simple and Efficient Pixel-Wise Training of Photometric Stereo Networks
• A-SDF: Learning Disentangled Signed Distance Functions for Articulated Shape Representation
  :star:code:house:project:tv:video
• ICON: Learning Regular Maps Through Inverse Consistency
• Video Geo-Localization Employing Geo-Temporal Feature Learning and GPS Trajectory Smoothing
  :star:code
• Kernel Methods in Hyperbolic Spaces
• Cross-Camera Convolutional Color Constancy
• BlockPlanner: City Block Generation with Vectorized Graph Representation
• A Machine Teaching Framework for Scalable Recognition
• Clothed Human Bodies
  • Dynamic Surface Function Networks for Clothed Human Bodies
    :star:code:house:project:tv:video
• 迁移学习
  • Fast and Efficient DNN Deployment via Deep Gaussian Transfer Learning
• Active Recognition(AR)
  • FLAR: A Unified Prototype Framework for Few-Sample Lifelong Active Recognition
• 3D摄影
  • SLIDE: Single Image 3D Photography with Soft Layering and Depth-aware Inpainting
    :open_mouth:oral:house:project:tv:video
• Sub-Bit Neural Networks: Learning To Compress and Accelerate Binary Neural Networks
  :star:code
• When Pigs Fly: Contextual Reasoning in Synthetic and Natural Scenes
  :star:code
• Physics-Enhanced Machine Learning for Virtual Fluorescence Microscopy
  :star:code
• Ground-truth or DAER: Selective Re-query of Secondary Information
  :star:code
• Can Shape Structure Features Improve Model Robustness Under Diverse Adversarial Settings?
• Joint Representation Learning and Novel Category Discovery on Single- and Multi-Modal Data
• Sparse Needlets for Lighting Estimation with Spherical Transport Loss
• Semantic Perturbations with Normalizing Flows for Improved Generalization
• Differentiable Surface Rendering via Non-Differentiable Sampling
• Towards Robustness of Deep Neural Networks via Regularization
• Objects as Cameras: Estimating High-Frequency Illumination from Shadows
• Inference of Black Hole Fluid-Dynamics From Sparse Interferometric Measurements
• Removing the Bias of Integral Pose Regression
• A Light Stage on Every Desk
  :house:project
• Multi-Level Curriculum for Training a Distortion-Aware Barrel Distortion Rectification Model
• Generic Event Boundary Detection: A Benchmark for Event Segmentation
• Extreme Structure from Motion for Indoor Panoramas without Visual Overlaps
  :star:code
• Continual Prototype Evolution: Learning Online from Non-Stationary Data Streams
  :star:code
• VaPiD: A Rapid Vanishing Point Detector via Learned Optimizers
• Multimodal Co-Attention Transformer for Survival Prediction in Gigapixel Whole Slide Images
  :star:code
• Efficient and Differentiable Shadow Computation for Inverse Problems
• Minimal Cases for Computing the Generalized Relative Pose using Affine Correspondences
• Radial Distortion Invariant Factorization for Structure from Motion
• LaLaLoc: Latent Layout Localisation in Dynamic, Unvisited Environments
• Transforms Based Tensor Robust PCA: Corrupted Low-Rank Tensors Recovery via Convex Optimization
• Synchronization of Group-labelled Multi-graphs
• Robust Watermarking for Deep Neural Networks via Bi-Level Optimization
• CrossNorm and SelfNorm for Generalization under Distribution Shifts
  :star:code
• Learning Temporal Dynamics from Cycles in Narrated Video
  :house:project
• von Mises-Fisher Loss: An Exploration of Embedding Geometries for Supervised Learning
• Multiple Heads are Better than One: Few-shot Font Generation with Multiple Localized Experts
  :star:code
• Me-Momentum: Extracting Hard Confident Examples From Noisily Labeled Data
  :star:code
• ProFlip: Targeted Trojan Attack with Progressive Bit Flips
• Attention Is Not Enough: Mitigating the Distribution Discrepancy in Asynchronous Multimodal Sequence Fusion
• AdvRush: Searching for Adversarially Robust Neural Architectures
• Improving robustness against common corruptions with frequency biased models
• UASNet: Uncertainty Adaptive Sampling Network for Deep Stereo Matching
• Glimpse-Attend-and-Explore: Self-Attention for Active Visual Exploration
  :star:code
• Field Convolutions for Surface CNNs
  :open_mouth:oral:star:code
• SIMstack: A Generative Shape and Instance Model for Unordered Object Stacks
• Learning Icosahedral Spherical Probability Map Based on Bingham Mixture Model for Vanishing Point Estimation
• Incorporating Learnable Membrane Time Constant to Enhance Learning of Spiking Neural Networks
  :star:code
• Real-Time Vanishing Point Detector Integrating Under-Parameterized RANSAC and Hough Transform
• Low-Rank Tensor Completion by Approximating the Tensor Average Rank
• Rotation Averaging in a Split Second: A Primal-Dual Method and a Closed-Form for Cycle Graphs
  :star:code
• Effectively Leveraging Attributes for Visual Similarity
  :star:code
• Localized Simple Multiple Kernel K-means
  :star:code
• SmartShadow: Artistic Shadow Drawing Tool for Line Drawings
• PT-CapsNet: A Novel Prediction-Tuning Capsule Network Suitable for Deeper Architectures
  :star:code
• Generalized Shuffled Linear Regression
  :star:code
• The Animation Transformer: Visual Correspondence via Segment Matching
• Weak Adaptation Learning: Addressing Cross-Domain Data Insufficiency With Weak Annotator
• Building-GAN: Graph-Conditioned Architectural Volumetric Design Generation
• Procrustean Training for Imbalanced Deep Learning