Analysis of Voice for Parkinson’s Disease Persons Using Dynamic Time Warping Technique
This paper presents a dynamic time warping (DTW) technique-based analysis of voice for distinguishing Parkinson’s disease (PD) persons from healthy persons. Mel frequency cepstral coefficient (MFCC) algorithm with MATLAB coding has been used to process voice samples. MFCC is converted into vector using MATLAB. DTW is useful for matching of voice samples. DTW-based matching percentage between PD-affected persons is 80.2163 %, whereas it is 72.2588 % between healthy persons. First coefficient of MFCC shows large values in case of PD-affected persons.
KeywordsAnalysis MFCC Dynamic time warping Parkinson’s disease Matching Voice
- 1.Lee, S.M., Fang, S.H., Hung, J.W., Lee, L.S.: Improved MFCC feature extraction by PCA-optimized filter bank for speech recognition. In: IEEE Workshop on Automatic Speech Recognition and Understanding, pp. 49–52 (2001) Google Scholar
- 2.Hossan, MA., Memon, S., Gregory, M.A.: A novel approach for MFCC feature extraction. In: 4th international conference on signal processing and communication systems (ICSPCS), pp. 1–5 (2010)Google Scholar
- 3.Bhalke, D.G., Rama Rao, C.B., Bormane, D.S.: Dynamic time warping technique for musical instrument recognition for isolated notes. In: Proceedings of ICETECT, pp. 768–771 (2011)Google Scholar
- 5.Giorgino, T.: Computing and visualizing dynamic time warping alignments in R: the dtw package. J. Stat. Softw. 31(7), 1–24 (2009)Google Scholar
- 6.Muller, M.: Information retrieval for music and motion, XVI, 318p. 136 illus. 39 in color., Hardcover. ISBN: 978-3-540-74047-6. http://www.springer.com/978-3-540-74047-6 (2007)