DETERMINATION OF CHANNEL-INDEPENDENT INFORMATION INDICATORS
Speech Technology Ltd., Minsk; Director
A. V. Tkachenia
Speech Technology Ltd., Minsk; Junior Scientist
M. V. Khitrov
STC Ltd., St. Petersburg; ITMO University, Department of Speech Information Systems; General Director; Head of the Department
Abstract. Information indicators of speech are analyzed for creation of channel-independent feature space aimed at improvement of speaker recognition system efficiency. For the problem of determination of similarity between several audio recordings, the optimal set of channel-independent information feature vectors is determined experimentally with the use of dynamic time warping.
Keywords: speech analysis, machine learning, feature selection, mel-frequency cepstral coefficients, dynamic time warping