This is the repository of the paper: "Features Understanding in 3D CNNs for Actions Recognition in Video" that has been accepted at The tenth International Conference on Image Processing Theory, Tools and Applications IPTA 2020, endorsed by IEEE.
Due to legal reasons, we have not uploaded the code of our proposed CNN Visualization Method yet. To understand and implement the method, please go through our paper.
Visualization of different classes of the experimented dataset is available at:
https://youtu.be/2yrG4vKxRTA
Thank you.