An ANN-based Method for Detecting Vocal Fold Pathology

TL;DR

This paper proposes a new feature extraction method using wavelet packet decomposition and MFCCs, combined with PCA for feature reduction, to improve vocal fold pathology detection using an ANN classifier.

Contribution

It introduces a novel feature vector combining wavelet packet decomposition and MFCCs, and evaluates its effectiveness with PCA and ANN for vocal fold pathology diagnosis.

Findings

Proposed feature vector improves classification accuracy.

PCA effectively reduces feature dimensionality.

ANN classifier achieves promising results with the new features.

Abstract

There are different algorithms for vocal fold pathology diagnosis. These algorithms usually have three stages which are Feature Extraction, Feature Reduction and Classification. While the third stage implies a choice of a variety of machine learning methods, the first and second stages play a critical role in performance and accuracy of the classification system. In this paper we present initial study of feature extraction and feature reduction in the task of vocal fold pathology diagnosis. A new type of feature vector, based on wavelet packet decomposition and Mel-Frequency-Cepstral-Coefficients (MFCCs), is proposed. Also Principal Component Analysis (PCA) is used for feature reduction. An Artificial Neural Network is used as a classifier for evaluating the performance of our proposed method.