On the Mutual Information between Source and Filter Contributions for Voice Pathology Detection

TL;DR

This paper explores the use of mutual information to evaluate features derived from speech and glottal signals for automatic voice pathology detection, highlighting their discriminative power and redundancy.

Contribution

It introduces a mutual information-based assessment of features from speech, glottal, and prosody signals for improved voice pathology detection.

Findings

Features related to the glottal source are highly informative.

Mutual information helps identify complementary features.

Analysis guides feature selection for better detection accuracy.

Abstract

This paper addresses the problem of automatic detection of voice pathologies directly from the speech signal. For this, we investigate the use of the glottal source estimation as a means to detect voice disorders. Three sets of features are proposed, depending on whether they are related to the speech or the glottal signal, or to prosody. The relevancy of these features is assessed through mutual information-based measures. This allows an intuitive interpretation in terms of discrimation power and redundancy between the features, independently of any subsequent classifier. It is discussed which characteristics are interestingly informative or complementary for detecting voice pathologies.