To detect the words spoken by a person based on his lip movements only, we employ two different approaches. One using Convolutional Neural Networks and other using Hidden Markov Model.
sindhura-pv/lip-reading
In this project, visual speech recognition has been attempted using 2 major machine learning techniques namely CNN and HMM. We also compare the efficiencies of Character and Word based CNN models. Miracl-VC1 Dataset was used to train all the models
Python