Tone Recognition Using Lifters and CTC

TL;DR

This paper introduces a novel neural network-based approach combining cepstrogram features and CTC for tone recognition in continuous speech, demonstrating superior performance on Mandarin Chinese speech data.

Contribution

It proposes a new method integrating cepstrogram features with CTC for improved tone recognition in tonal languages.

Findings

Outperforms existing techniques in tone error rate

Effective on Mandarin Chinese speech corpus AISHELL-1

Demonstrates the viability of CNN-CTC approach for tonal speech recognition

Abstract

In this paper, we present a new method for recognizing tones in continuous speech for tonal languages. The method works by converting the speech signal to a cepstrogram, extracting a sequence of cepstral features using a convolutional neural network, and predicting the underlying sequence of tones using a connectionist temporal classification (CTC) network. The performance of the proposed method is evaluated on a freely available Mandarin Chinese speech corpus, AISHELL-1, and is shown to outperform the existing techniques in the literature in terms of tone error rate (TER).