A Cycle-GAN Approach to Model Natural Perturbations in Speech for ASR Applications

TL;DR

This paper introduces a CycleGAN-based front-end to transform naturally perturbed speech into normal speech, significantly enhancing ASR robustness against emotional and physical speaker states.

Contribution

The paper presents a novel CycleGAN approach trained on non-parallel data to improve speech recognition in emotionally and physically perturbed speech conditions.

Findings

ASR performance improves with CycleGAN-processed speech

Effective transformation of laughter and creaky speech

Visualization confirms feature similarity post-transformation

Abstract

Naturally introduced perturbations in audio signal, caused by emotional and physical states of the speaker, can significantly degrade the performance of Automatic Speech Recognition (ASR) systems. In this paper, we propose a front-end based on Cycle-Consistent Generative Adversarial Network (CycleGAN) which transforms naturally perturbed speech into normal speech, and hence improves the robustness of an ASR system. The CycleGAN model is trained on non-parallel examples of perturbed and normal speech. Experiments on spontaneous laughter-speech and creaky-speech datasets show that the performance of four different ASR systems improve by using speech obtained from CycleGAN based front-end, as compared to directly using the original perturbed speech. Visualization of the features of the laughter perturbed speech and those generated by the proposed front-end further demonstrates the effectiveness of our approach.