This is the project content of frame interpolation. Now we are implementing the paper Softmax Splatting for Video Frame Interpolation [1]. We'll try various techniques to improve the quality of output frames.
The following packages need to be installed:
pip install cupy-cuda101
if your cuda version is v10.1, please use the version corresponding to your cuda version
pip install opencv-contrib-python
pip install torch
pip install tensorflow
if version is higher than 1.15, it may fails.
Before generating frames, you need to check whether the flow files exist. If not, you can cd PWCNet
, and then execute python predict.py --task type_of_task
. After that, the frames are generated by execute python main.py --task type_of_task
. Finally, you can simply execute bash eval.sh
to check the quality of generated frames.
[1] @inproceedings{Niklaus_CVPR_2020,
author = {Simon Niklaus and Feng Liu},
title = {Softmax Splatting for Video Frame Interpolation},
booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},
year = {2020}
}