Video Action Recognition with PyTorch

Rough PyTorch implementation of "Action Recognition in Video Sequences using Deep Bi-directional LSTM with CNN Features" (Amin Ullah, et al.)

Disclaimer

This implementation is rough. Although this model should work for video action recognition, here should be many errors and/or differences with the original paper.

Paper citation

Amin Ullah, Jamil Ahmad, Khan Muhammad, Muhammad Sajjad, and Sung Wook Baik. "Action recognition in video sequences using deep bi-directional LSTM with CNN features." IEEE Access 6 (2017): 1155-1166. doi: 10.1109/ACCESS.2017.2778011.

More accurate implementation

You can find tensorflow implementation written by the paper author in Aminullah6264/BidirectionalLSTM.

How to use

Modify options in main.py first.

If using Google Colab, mount Google Drive before running.

python main.py

Requirements

pip install torch===1.7.1+cu110 torchvision===0.8.2+cu110 torchaudio===0.7.2 -f https://download.pytorch.org/whl/torch_stable.html

pip install numpy requests av

Working period

20 Nov 2020 - 19 Jan 2021 (62 days)

License and community improvements

• MIT License
• If you are interested in fixing errors and/or differences with the original paper, making a pull request is always sincerely welcome.