Augmenting Images for ASR and TTS through Single-loop and Dual-loop Multimodal Chain Framework

TL;DR

This paper introduces a novel multimodal chain framework that uses image generation to augment data for improving ASR and TTS systems, demonstrating effectiveness with natural multispeaker speech data.

Contribution

It proposes a revamped multimodal chain framework with image generation and explores single-loop and dual-loop architectures for data augmentation in ASR and TTS.

Findings

Both frameworks improved ASR and TTS performance using image-only data.

The approach works with multispeaker natural speech data.

Image generation effectively augments data for speech tasks.

Abstract

Previous research has proposed a machine speech chain to enable automatic speech recognition (ASR) and text-to-speech synthesis (TTS) to assist each other in semi-supervised learning and to avoid the need for a large amount of paired speech and text data. However, that framework still requires a large amount of unpaired (speech or text) data. A prototype multimodal machine chain was then explored to further reduce the need for a large amount of unpaired data, which could improve ASR or TTS even when no more speech or text data were available. Unfortunately, this framework relied on the image retrieval (IR) model, and thus it was limited to handling only those images that were already known during training. Furthermore, the performance of this framework was only investigated with single-speaker artificial speech data. In this study, we revamp the multimodal machine chain framework with image generation (IG) and investigate the possibility of augmenting image data for ASR and TTS using single-loop and dual-loop architectures on multispeaker natural speech data. Experimental results revealed that both single-loop and dual-loop multimodal chain frameworks enabled ASR and TTS to improve their performance using an image-only dataset.