Voiceprint identification method based on multi-type combination characteristic parameters

Abstract

The invention discloses a voiceprint identification method based on multi-type combination characteristic parameters, and the method comprises the following steps: the collecting and inputting of voice signals; the preprocessing of voice signals; the extracting of combination characteristic parameters of the voice signals: the extracting of a multi-dimensional feature vector of MFCC, LPCC, delta MFCC, delta LPCC, energy, first difference of energy and GFCC characteristic parameters; the screening of the multi-dimensional feature parameters through a genetic algorithm; the obtaining a voice model of a speaker through the introduction of a universal background model UBM; and the recognition of test voice through a GMM-UBM model at last. Compared with a voiceprint identification method for single voice signal characteristic parameter, the method effectively improves the recognition accuracy and system stability of voiceprint recognition.

Description

technical field [0001] The invention relates to the field of voice signal processing, in particular to a voiceprint recognition method based on multi-type combination feature parameters. Background technique [0002] Under the premise of today's information age, identification technology, one of the important components of information security, has brought new challenges. Due to the limitations of the algorithm and the rise of hardware and software decryption technology, the traditional password identification has shown its disadvantages. As one of the new technologies of identification, voiceprint recognition technology has attracted more and more attention because of its unique advantages of convenience, economy and accuracy. [0003] Voiceprint recognition is to extract the speaker's personality characteristics from a segment of the speaker's voice, and through the analysis and identification of personal characteristics, so as to achieve the purpose of identifying or con...

AI Technical Summary

Problems solved by technology

[0007] The present application provides a voiceprint recognition method based on multi-type combination feature parameters, which includes the following steps: voice signal collection and input, voice signal preprocessing, and voice signal combination feature parameter extraction: that is, extracting MFCC, LPCC, △MFCC, △ LPCC, energy, the first-order difference of energy, and GFCC feature parameters together form a multidimensional feature vector, use genetic algorithm to screen the multidimensional feature parameters, introduce the general background model UBM training to get the speaker's voice model, and finally use the GMM-UBM model to test Speech recognition to solve the technical problems in the prior art that the accuracy of voiceprint recognition using a single voice parameter is not high and the voiceprint recognition system is unstable

Images