Pytorch implementation of tvNet by LijieFan
The original implementation is by tensorflow and most of the code this Pytorch implementation is adapted from that one
Note that the code of optic flow visualization is from FlowNetPytorch by ClementPinard, the visualization scheme seems different from the one provided in the original implementation.
| frame1 | frame2 | prediction |
|---|---|---|
 |
 |
 |
@inproceedings{fan2018end,
title={End-to-End Learning of Motion Representation for Video Understanding},
author={Fan, Lijie and Huang, Wenbing and Gan, Chuang and Ermon, Stefano and Gong, Boqing and Huang, Junzhou},
booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
pages={},
year={2018}
}