Convolutional Neural Networks

September 17, 2025 ยท View on GitHub

ModelDescriptionVenueYear
SRCNN3-layer convolutional networkECCV2014
CSCNsparse coding and deep learningICCV2015
FSRCNNtransposed convolutionECCV2016
ESPCNefficient sub-pixel convolutionCVPR2016
VDSRdeep networksCVPR2016
DRCNrecursive CNNCVPR2016
EDSRremoving the Batch Normalization layer in the residual blockCVPR2017
LapSRNprogressively reconstruct the sub-bandCVPR2017
DRRNDeep Recursive Residual NetworkCVPR2017
SRDenseNetDenseNetICCV2017
MemNetmemory block, consisting of a recursive unit and a gate unitICCV2017
RDNcombining residuals and dense connectionsCVPR2018
DSRNDual-State Recurrent NetworksCVPR2018
SRMDNFa Single Convolutional Super-Resolution Network for Multiple DegradationsCVPR2018
IDNinformation distillation blocksCVPR2018
DBPNiterative up- and down- sampling layersCVPR2018
ZSSRexamples extracted solely from the input image itselfCVPR2018
CARNcascading mechanism upon a residual networkECCV2018
RCANResidual Channel Attention NetworksECCV2018
MSRNMulti-scale Residual NetworkECCV2018
IMDNcascaded information multi-distillation blocksACMMM2019
SRFBNfeedback networkCVPR2019
OISRordinary differential equation (ODE)-inspired design schemeCVPR2019
META-SRarbitrary SRCVPR2019
SANsecond-order attention networkCVPR2019
RNANlocal and non-local attention blocksICLR2019
EBRNEmbedded Block Residual NetworkICCV2019
SCNpyramid representation via bilinear down-scalingAAAI2020
MZSRMeta-Transfer LearningCVPR2020
DRNdual regression scheme to reduce space of the possible functionsCVPR2020
cutBlurdata preprocessingCVPR2020
CSNLNCross-Scale Non-Local Attention and Exhaustive Self-Exemplars MiningCVPR2020
RFANetresidual feature aggregationCVPR2020
HANlayer attention module (LAM) and a channel-spatial attention module (CSAM)ECCV2020
IRNInvertible Rescaling NetECCV2020
LatticeNettwo butterfly structures are applied to combine two RBsECCV2020
ClassSRcombined classification and SR in a unified frameworkCVPR2021
FKPnormalizing flow-based kernel prior (FKP) for kernel modelingCVPR2021
SRWraparbitrary image transformationCVPR2021
NLSAdynamic sparse attention patternCVPR2021
LUTprecomputed HR output values from the tableCVPR2021
SMSRsparse masks, sparse convolutionCVPR2021
AdderSRadder networksCVPR2021
FADfrequency-aware dynamic network, DCTICCV2021
DFSAlearning of high frequency features both locally and globallyICCV2021
ArbRCANscale-arbitrary SRICCV2021
NAPSNAS, pruningICCV2021
CRANContext Reasoning Attention NetworkICCV2021
LIIFimplicit neural representationCVPR2021
DCLSdeconvolution module, channel-wise deblurringCVPR2022
SLSpruning methodCVPR2022
LTEdominant-frequency estimatorCVPR2022
SESRover-parameterizeMLS2022
SPLUTSeries-Parallel Lookup TablesECCV2022
HPUNefficient and effective downsampling modulearxiv2022
CMOSmulti modalities combinationCVPR2023
CABMbit mappingCVPR2023
LINFnormalizing flowCVPR2023
ETDSlightweight srCVPR2023
MSSRscalable srCVPR2023
SUDFfrequency-based srCVPR2023

Transformer-based Methods

ModelDescriptionPublicationYear
IPTpre-trained modelCVPR2021
SwinIRbased on the Swin TransformerICCVW2021
CATrectangle transformerNIPS2022
ESRTLightweight TransformerCVPR2022
ELANshift convolution, group-wise self-attentionECCV2022
LBNetLightweight Bimodal NetworkIJCAI2022
HATHybrid Attention TransformerCVPR2023
EQSRArbitrary-ScaleCVPR2023
CiaoSRArbitrary-ScaleCVPR2023
NGrammodification of windowCVPR2023
Omni-SRspatial and channel attentionCVPR2023
CLITLocal Implicit TransformerCVPR2023