Visual Speech Recognition Using PCA Networks and LSTMs in a Tandem GMM-HMM System

TL;DR

This paper presents a novel visual speech recognition system combining PCA-based convolutional networks and LSTMs within a tandem GMM-HMM framework, improving phrase recognition accuracy on the OuluVS2 database.

Contribution

It introduces a new method integrating PCA networks and LSTMs for visual speech recognition, outperforming baseline techniques on a standard dataset.

Findings

Achieved 79% cross-validation accuracy, surpassing the baseline of 74%.

Improved phrase recognition accuracy on the OuluVS2 database.

Demonstrated effectiveness of the combined PCA-LSTM approach in noisy or audio-limited scenarios.

Abstract

Automatic visual speech recognition is an interesting problem in pattern recognition especially when audio data is noisy or not readily available. It is also a very challenging task mainly because of the lower amount of information in the visual articulations compared to the audible utterance. In this work, principle component analysis is applied to the image patches - extracted from the video data - to learn the weights of a two-stage convolutional network. Block histograms are then extracted as the unsupervised learning features. These features are employed to learn a recurrent neural network with a set of long short-term memory cells to obtain spatiotemporal features. Finally, the obtained features are used in a tandem GMM-HMM system for speech recognition. Our results show that the proposed method has outperformed the baseline techniques applied to the OuluVS2 audiovisual database for phrase recognition with the frontal view cross-validation and testing sentence correctness reaching 79% and 73%, respectively, as compared to the baseline of 74% on cross-validation.