Speech signal feature parameters extraction algorithm based on PCNN for isolated word recognition

Abstract

In the isolated word speech recognition system, the extraction and matching of the characteristic parameters is the key link. This paper introduces a new feature parameter extracting methods on basis of Pulse Coupled Neural Network (PCNN) for the recognition system. By means of the visibility of speech spectrogram, the PCNN is used to extract the time series and entropy series from the spectrogram of words. Finally, by means of DTW algorithm to accomplish the task of isolated word recognition, the simulation results demonstrate the feasibility and effectiveness of the proposed algorithm.