A Hybrid of Genetic Algorithm and Gaussian Mixture Model for Features Reduction and Detection of VocalFold Pathology

Acoustic analysis is a proper method in vocal fold pathology diagnosis so that itcan complement and in some cases replace the other invasive, based on direct vocalfold observation, methods. There are different approaches and algorithms for vocalfold pathology diagnosis. These algorithms usually have three stages which areFeature Extraction, Feature Reduction and Classification. In this paper initial studyof feature extraction and feature reduction in the task of vocal fold pathologydiagnosis has been presented. A new type of feature vector, based on wavelet packetdecomposition and Mel-Frequency-Cepstral-Coefficients (MFCCs), is proposed.Also a new GA-based method for feature reduction stage is proposed and comparedwith conventional methods such as Principal Component Analysis (PCA). GaussianMixture Model (GMM) is used as a classifier for evaluating the performance of theproposed method. The results show the priority of the proposed method incomparison with current methods.

Keywords:

Vocal Fold Pathology Diagnosis , Wavelet Packet Decomposition(WPD) , Mel-Frequency-Cepstral-Coefficient (MFCC) , Principal Component Analysis(PCA) , Genetic Algorithm (GA) , Gaussian Mixture Model (GMM)