Home

Awesome

EAVSR (CVPRW 2023)

PyTorch implementation of Benchmark Dataset and Effective Inter-Frame Alignment for Real-World Video Super-Resolution

1. MVSR4× Datasets

<p align="center"><img src="./fig/dataset.png" width="95%"></p> <p align="center">Some samples of proposed MVSR4× dataset.</p>

2. Framework

<p align="center"><img src="./fig/multiadastn.png" width="95%"></p> <p align="center">Overall pipeline of proposed MultiAdaSTN.</p>

3. Quick Start

3.1 Training

3.3 Inference

3.4 Note

4. Citation

If you find it useful in your research, please consider citing:

@InProceedings{Wang_2023_CVPR,
    author    = {Wang, Ruohao and Liu, Xiaohui and Zhang, Zhilu and Wu, Xiaohe and Feng, Chun-Mei and Zhang, Lei and Zuo, Wangmeng},
    title     = {Benchmark Dataset and Effective Inter-Frame Alignment for Real-World Video Super-Resolution},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
    month     = {June},
    year      = {2023},
    pages     = {1168-1177}
}

5. Acknowledgement

This repo is built upon the framework of CycleGAN, and we borrow some code from RealVSR and BasicVSR++, thanks for their excellent work!