A speech endpoint detector based on eigenspace-energy-entropy
Wang Bingxi, Pla Information · Journal of China Institute of Communications · 2003
We perform the eigen-decomposition to the covariance matrix of speech signal. Using the probability density function of the signal抯 eigenspace-energy evaluated by the eigenvalues, we propose the conception of eigenspace-energy-entropy. The experiments in different noise backgrounds show that high endpoint detection accuracy can be obtained by this entropy.