This is an implementation of LipNet for CSCE 629 - Neural Networks at Texas A&M University. Reference - https://github.com/rizkiarm/LipNet Research Paper - https://arxiv.org/abs/1611.01599
The repository contains a small subset of GRID CORPUS to see how code for preprocessing, training and prediction works.
Run mouth_extract.py in the MouthExtract directory. The code loads videos from the Video folder and saves each video as 75 mouth-crop frames.
Run train.py in the Training directory. The code loads training data which is preprocessed.
Run GUI.py in the PredictGUI directory.