The Research Of Anti-Noise Speech Recognition Feature

Speech recognition has wide use in the field of communication and so on. Speech feature parameter extraction is an important part of the speech recognition system. The performance of the feature parameter influences the system's performance directly. And the environmental noise is a key factor of restricting the performance of the feature parameter. This paper's research object is extracting the speech feature parameter under the noise environment. Then it analyzes the auditory model of human, deeply analyzes and researches the traditional speech feature parameter MFCC, and proposes three improved MFCC parameters based on the wavelet transformation and the auditory characteristic of human. Besides, the paper also gives an improved method about the feature parameter ZCPA.Firstly, the thesis introduces the traditional speech feature coefficient MFCC, which is designed based on the auditory characteristic of human. And we can obtain high recognition rate in a quiet environment by using of the MFCC. However the recognition rate drops suddenly in the low SNR...