README.md

April 19, 2025 · View on GitHub

DeNVeR: Deformable Neural Vessel Representations for Unsupervised Video Vessel Segmentation

CVPR 2025

Dataset path

Download XACV from google drive
Replace xca_dataset with our XACV dataset.

XACV
├── CVAI-2828
│   ├── ground_truth
│   │   ├── CVAI-2828RAO2_CRA32
│   │   │   ├── 00056.png
│   │   │   └── 00065.png
│   │   └── CVAI-2828RAO2_CRA32CATH
│   │       ├── 00056.png
│   │       └── 00065.png
│   └── images
│       └── CVAI-2828RAO2_CRA32
│           ├── 00000.png
│           ├── 00001.png
│           ├── 00002.png
│           ├── 00003.png
│           ├── 00004.png
│           ├── ...
├── CVAI-2829
├── ...

Requirements

Python versionn = 3.9

conda env create -f environment.yml

Download RAFT checkpoint

cd RAFT
./download_models.sh

Test-time training

python main.py -d {seq_name}

e.g.

python main.py -d CVAI-2828RAO2_CRA32

Batch running

python launch.py -f ./job_specs/vessel.txt --gpus {gpu_ids}

Evaluation

cd eval
python gt_filename.py -d {exp_name}
python eval.py

exp_name is the last directory name in the dir path specified in confs/config.yaml. e.g.

python gt_filename.py -d init_model

Citation

Please cite us if our work is useful for your research.

@inproceedings{wu2025denver,
  title={DeNVeR: Deformable Neural Vessel Representations for Unsupervised Video Vessel Segmentation},
  author={Wu, Chun-Hung and Chen, Shih-Hong and Hu, Chih Yao and Wu, Hsin-Yu and Chen, Kai-Hsin and Chen, Yu-You and Su, Chih-Hai and Lee, Chih-Kuo and Liu, Yu-Lun},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2025}
}

This research was funded by the National Science and Technology Council, Taiwan, under Grants NSTC 112-2222-E-A49-004-MY2. The authors are grateful to Google, NVIDIA, and MediaTek Inc. for generous donations. Yu-Lun Liu acknowledges the Yushan Young Fellow Program by the MOE in Taiwan.