Awesome

Collaborative Feedback Discriminative Propagation for Video Super-Resolution

[Paper (arxiv)]   [Project Page]

Hao Li, Xiang Chen, Jiangxin Dong, Jinhui Tang, Jinshan Pan <br> IMAG Lab, Nanjing University of Science and Technology

CFDVSR mainly consists of a propagation module with the discriminative alignment correction (DAC) and a collaborative feedback propagation (CFP) module. The DAC uses shallow features ft to explore more details information after feature warping, which corrects the aligned features for propagation. The Feedback ConvGRU and gated collaborate feed-forward block (GCFB) in CFP bring more temporal interactions between different timestep featuresfrom forward and backward propagation simultaneously.

News

• [2024-04-09] The paper is available here.

To do

• Release training code
• Release testing code and pre-trained models

Results

• Quantitative Results

  <img width="800" src="./assets/quantitative_results.png">

• Qualitative Results

  <img width="800" src="./assets/qualitative_1.png">

• Video Demo

https://github.com/House-Leo/CFDVSR/assets/47940711/ed207ca3-ff43-4514-a436-df63df67e41c.mp4

https://github.com/House-Leo/CFDVSR/assets/47940711/9ad3d6fe-d462-4e68-a915-3492406420dd.mp4

Citation

If this work is helpful for your research, please consider citing the following BibTeX entry.

@article{li2024cfdvsr,
    title={Collaborative Feedback Discriminative Propagation for Video Super-Resolution},
    author={Li, Hao and Chen, Xiang and Dong, Jiangxin and Tang, Jinhui and Pan, Jinshan},
    journal={arXiv preprint arXiv:2404.04745},
    year={2024}
 }

Acknowledgement

The README template is borrowed from SAFMN. Thanks for the awesome work.

Contact

If you have any questions, please feel free to reach me out at <a href="mailto:haoli@njust.edu.cn">haoli@njust.edu.cn</a>.