Deep Learning Models for Automated Lip Reading (ALR)
Run download_grid.py. It will create a grid_videos folder in your home directory and download all the videos to it.
Then run scripts/extract_mouth_region.py. It will create a grid_images folder in your home directory and populate it with cropped images.
Although pretrained weights and config files are present, there is no need to use them.
Model architectures are in /model_archs.
Training scripts are in /train_scripts. You can create configs in order to store hyperparameters.