Estimating speech from lip dynamics

TL;DR

This paper presents a lip reading system that predicts spoken words from lip movements alone, using Hidden Markov Models to classify visemes and phonemes without audio input.

Contribution

It introduces a limited lip reading algorithm for English that operates solely on visual lip data and classifies speech units for word prediction.

Findings

Achieved word prediction accuracy on the GRID corpus

Effective classification of visemes and phonemes from lip movements

Demonstrated feasibility of audio-free speech recognition

Abstract

The goal of this project is to develop a limited lip reading algorithm for a subset of the English language. We consider a scenario in which no audio information is available. The raw video is processed and the position of the lips in each frame is extracted. We then prepare the lip data for processing and classify the lips into visemes and phonemes. Hidden Markov Models are used to predict the words the speaker is saying based on the sequences of classified phonemes and visemes. The GRID audiovisual sentence corpus [10][11] database is used for our study.