The project tries to solve a speaker diarization problem using audio features, face recognition and video feature extraction from face image, mouth tracking.

https://github.com/valiakon/MultimodalAnalysis_SpeakerDiarization