Deep Bayesian Video Frame Interpolation (ECCV2022)

[Paper&supp], [Demo], [Presentation]

1. Requirements

1. cuda 9.0, cudnn7.6.5

2. python 3.6.9

3. pytorch 1.8.1

4. numpy 1.17.2

5. cupy-90

6. tqdm

7. gcc 5.4.0

8. cmake 3.16.0

9. opencv_contrib_python

10. Apex

11. For distributed training with multi-gpus on cluster: slurm 15.08.11

2. How to use

For 8x interpolation

For 2x interpolation

3. Citation

@inproceedings{DBVI2022,
  title={Deep Bayesian Video Frame Interpolation},
  author={Yu, Zhiyang and Zhang, Yu and Xiang, Xujie and Zou, Dongqing and Chen, Xijun and Ren, Jimmy S},
  booktitle={European Conference on Computer Vision},
  pages={144--160},
  year={2022},
  organization={Springer}
}

4. Reference code base

[opencv_torchvision], [ESRGAN], [CAM-Net], [SoftSplat], [DeepView], [FLAVR], [superSlomo], [QVI]