Spectro-Temporal Deep Features for Disordered Speech Assessment and Recognition

TL;DR

This paper introduces spectro-temporal deep features derived from SVD decomposition of speech spectra to improve disordered speech recognition and assessment, demonstrating significant WER reductions on the UASpeech corpus.

Contribution

It proposes novel spectro-temporal subspace basis embedding deep features for disordered speech recognition and speaker adaptation, outperforming traditional i-Vector methods.

Findings

Achieved up to 8.6% relative WER reduction over baseline

Consistent improvements with data augmentation and LHUC adaptation

Final system attained 25.6% WER on UASpeech test set

Abstract

Automatic recognition of disordered speech remains a highly challenging task to date. Sources of variability commonly found in normal speech including accent, age or gender, when further compounded with the underlying causes of speech impairment and varying severity levels, create large diversity among speakers. To this end, speaker adaptation techniques play a vital role in current speech recognition systems. Motivated by the spectro-temporal level differences between disordered and normal speech that systematically manifest in articulatory imprecision, decreased volume and clarity, slower speaking rates and increased dysfluencies, novel spectro-temporal subspace basis embedding deep features derived by SVD decomposition of speech spectrum are proposed to facilitate both accurate speech intelligibility assessment and auxiliary feature based speaker adaptation of state-of-the-art hybrid DNN and end-to-end disordered speech recognition systems. Experiments conducted on the UASpeech corpus suggest the proposed spectro-temporal deep feature adapted systems consistently outperformed baseline i-Vector adaptation by up to 2.63% absolute (8.6% relative) reduction in word error rate (WER) with or without data augmentation. Learning hidden unit contribution (LHUC) based speaker adaptation was further applied. The final speaker adapted system using the proposed spectral basis embedding features gave an overall WER of 25.6% on the UASpeech test set of 16 dysarthric speakers