End-to-end Audiovisual Speech Recognition

TL;DR

This paper introduces a novel end-to-end audiovisual speech recognition model that learns directly from raw images and audio signals, outperforming audio-only models especially in noisy environments.

Contribution

It presents the first audiovisual fusion model that simultaneously learns feature extraction and recognition from raw pixels and waveforms using residual networks and BGRUs.

Findings

Outperforms audio-only models in noisy conditions

Achieves slight improvement in clean audio recognition

Significantly better performance under high noise levels

Abstract

Several end-to-end deep learning approaches have been recently presented which extract either audio or visual features from the input images or audio signals and perform speech recognition. However, research on end-to-end audiovisual models is very limited. In this work, we present an end-to-end audiovisual model based on residual networks and Bidirectional Gated Recurrent Units (BGRUs). To the best of our knowledge, this is the first audiovisual fusion model which simultaneously learns to extract features directly from the image pixels and audio waveforms and performs within-context word recognition on a large publicly available dataset (LRW). The model consists of two streams, one for each modality, which extract features directly from mouth regions and raw waveforms. The temporal dynamics in each stream/modality are modeled by a 2-layer BGRU and the fusion of multiple streams/modalities takes place via another 2-layer BGRU. A slight improvement in the classification rate over an end-to-end audio-only and MFCC-based model is reported in clean audio conditions and low levels of noise. In presence of high levels of noise, the end-to-end audiovisual model significantly outperforms both audio-only models.