Multistream neural architectures for cued-speech recognition using a pre-trained visual feature extractor and constrained CTC decoding

TL;DR

This paper introduces a multistream neural network system utilizing pre-trained visual feature extractors and constrained CTC decoding for improved automatic recognition of Cued Speech, achieving competitive phonetic accuracy on the French CS dataset.

Contribution

It presents a novel combination of pre-trained visual feature extraction and multistream RNN with CTC for Cued Speech recognition, outperforming previous CNN-HMM methods.

Findings

Phonetic decoding accuracy of 70.88% on CSF18 dataset.

Outperforms previous CNN-HMM decoder.

Competitive with more complex baselines.

Abstract

This paper proposes a simple and effective approach for automatic recognition of Cued Speech (CS), a visual communication tool that helps people with hearing impairment to understand spoken language with the help of hand gestures that can uniquely identify the uttered phonemes in complement to lipreading. The proposed approach is based on a pre-trained hand and lips tracker used for visual feature extraction and a phonetic decoder based on a multistream recurrent neural network trained with connectionist temporal classification loss and combined with a pronunciation lexicon. The proposed system is evaluated on an updated version of the French CS dataset CSF18 for which the phonetic transcription has been manually checked and corrected. With a decoding accuracy at the phonetic level of 70.88%, the proposed system outperforms our previous CNN-HMM decoder and competes with more complex baselines.