Can DNNs Learn to Lipread Full Sentences?
This addresses the problem of automated lipreading for complex, unconstrained speech, offering a significant but incremental advance in visual speech recognition.
The paper tackled the challenge of lipreading full sentences by exploring state-of-the-art Deep Neural Network architectures, reporting a major improvement over Hidden Markov Model frameworks on the TCD-TIMIT dataset with 59 speakers and over 6000 words.
Finding visual features and suitable models for lipreading tasks that are more complex than a well-constrained vocabulary has proven challenging. This paper explores state-of-the-art Deep Neural Network architectures for lipreading based on a Sequence to Sequence Recurrent Neural Network. We report results for both hand-crafted and 2D/3D Convolutional Neural Network visual front-ends, online monotonic attention, and a joint Connectionist Temporal Classification-Sequence-to-Sequence loss. The system is evaluated on the publicly available TCD-TIMIT dataset, with 59 speakers and a vocabulary of over 6000 words. Results show a major improvement on a Hidden Markov Model framework. A fuller analysis of performance across visemes demonstrates that the network is not only learning the language model, but actually learning to lipread.