Distortion measure of spectrograms for classification of respiratory diseases

TL;DR

This paper introduces a novel feature extraction method from spectrogram surfaces of pulmonary sounds, combined with MFCCs, to improve automatic classification of respiratory diseases, outperforming baseline methods.

Contribution

The paper presents a new geometrical distortion-based feature extraction technique from spectrograms for respiratory disease classification, enhancing diagnostic accuracy.

Findings

Superior classification accuracy over baseline methods

Effective combination of geometrical features with MFCCs

Applicable to other one-dimensional signals in manifold representations

Abstract

A new method for the classification of respiratory diseases is presented. The method is based on a novel class of features, extracted from pulmonary sounds, by parameterizing their spectrograms that are represented as surfaces, and by utilizing geometrical distortions defined with reference to these surfaces. This method yields a set of highly descriptive features for the analysis of pulmonary sound recordings. Furthermore, by combining these features with Mel-frequency cepstral coefficients, we introduce a powerful model for the automatic diagnosis of common respiratory pathologies. Compared with baseline methods, our model achieves superior results for binary and multi-class classifications of common respiratory diseases. Our new approach to the classification of one-dimensional signals is applicable to other signals in the context of their representations in combined spaces or manifolds.