Study on speech endpoint detection algorithm based on wavelet energy entropy
Cao Yali, Gao Jing, Guang Yang · 2016
This paper describes a high quality voice activity detection using wavelet energy entropy. In this algorithm, the partitioning of the speech band into sub-bands is performed via a bank of the adaptive band-partitioning filters whose coefficients are derived from a wavelet tree structure. The adaptive band-partitioning models have been proposed to perform endpoint detections of isolated digit utterances spoken in the Language. We use the wavelet energy entropy to analyze each sub-band of speech signal. The results of the proposed algorithm on adaptive band-partitioning of signals and noise regions show a good performance of this method.