Voice processing technique for patients with stroke based on chao theory and surrogate data analysis
Jian Li · Journal of Zhejiang University(Engineering Science) · 2015
Human voices changing as a stroke result were analyzed.A method based on chaos and surrogate data method was proposed to analyze the voice signal from nonlinear aspect.The correlation dimension,the largest Lyapunov exponent of the voice signal and value of first minimum of mutual information function were calculated.There exist several fluctuations in voice when stroke patients speaking,which increases the correlation dimension of the voice signal.A new feature was proposed which is normalized variance quantity based on surrogate data method and the correlation dimension of voice.The voice analysis results of 17 patients and 20 healthy people show that the difference between the chaotic features of the normal sounds and that of the pathological sounds is significant.Results prove that the dysfunction in brain causes the disorder in voice.