Title: Disease detection using voice analysis: a review

Authors: Saloni; R.K. Sharma; A.K. Gupta

Addresses: Electronics and Communication Engineering Department, National Institute of Technology, Kurukshetra 136119, India ' Electronics and Communication Engineering Department, National Institute of Technology, Kurukshetra 136119, India ' Electronics and Communication Engineering Department, National Institute of Technology, Kurukshetra 136119, India

Abstract: Disease detection using voice analysis is a vital research topic in medical engineering. It is a reliable, efficient, economic and easy to use method. It also helps to detect the diseases at its earlier stage. Voice features are extracted using some DSP techniques. These features contain information on health of voice tract and of the organs cooperating in speech production. These features represent the particular voice and may be used to discriminate voice of healthy and unhealthy persons. The time domain analysis, spectrum analysis, cepstrum analysis, glottal waveform analysis are used to extract the voice features. These features are then classified into groups using various classification techniques like vector quantisation, dynamic time wrapping, support vector machine, Gaussian mixture model and artificial neural network. This paper described all these techniques in detail.

Keywords: human voice; medical engineering; vocal disease; voice analysis; feature extraction; feature classification; disease detection; voice features; vector quantisation; dynamic time wrapping; support vector machine; SVM; Gaussian mixture model; artificial neural networks; ANNs.

DOI: 10.1504/IJMEI.2014.063173

International Journal of Medical Engineering and Informatics, 2014 Vol.6 No.3, pp.189 - 209

Received: 14 May 2013
Accepted: 18 Nov 2013
Published online: 26 Jul 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article