Kwanss / PCLNet

Unsupervised Learning for Optical Flow Estimation Using Pyramid Convolution LSTM.

PCLNet

This repo contains source code for the paper:

Unsupervised Learning for Optical Flow Estimation Using Pyramid Convolution LSTM. In Proceedings of IEEE International Conference on Multimedia and Expo(ICME). 2019.

Framework

Acknowledgement

The implementation for Mean Structural Similarity (MSSIM) metric (models/ssim_module.py) is derived from: https://github.com/Po-Hsun-Su/pytorch-ssim.git

Citations

If you use PCLNet, please cite the following paper:

@InProceedings{PCLNet-icme2019,
  author    = {Shuosen Guan, Haoxin Li, Wei-Shi Zheng},
  title     = {Unsupervised Learning for Optical Flow Estimation Using Pyramid Convolution LSTM},
  booktitle = {Proceedings of IEEE International Conference on Multimedia and Expo(ICME)},
  year      = {2019},
}