Short Speech Speaker Recognition Method Based on Basis State Vector Weighting

Abstract

The invention provides a short voice speaker recognizing method based on base state vector weighting. The method comprises the following steps of: acquiring multiple voice data with text marks and training the multiple voice data to obtain a state-layer-clustered hidden markov model; decoding the data in speaker recognition according to the state-layer-clustered hidden markov model to obtain base state marks of the data; training a universal background model of the base state according to the base state marks of the data, and generating a base state mean super-vector and a base state weight super-vector from the model after MAP self-adaption; and according to the base state mean super-vector and the base state weight super-vector, implementing model training to a speaker and testing and estimating the identity of the short voice speaker. According to the method provided by the embodiment of the invention, fine modeling of the base state layer can be realized, and the problem of 'model hole' easily caused by the traditional method is overcome by effective weighting, so that the freedom of modeling is effectively lowered and simultaneously the recognition performance of the speaker is enhanced.

Description

technical field [0001] The invention relates to the technical field of speech recognition, in particular to a short speech speaker recognition method based on base state vector weighting. Background technique [0002] Speaker recognition technology refers to a biometric identification technology that uses a machine to automatically identify the speaker's identity information from the speech signal to be tested. This technology is widely used in voice-based speaker identification, public security criminal investigation, court evidence identification, national security and other fields. [0003] Common speaker recognition systems mainly include VQ (Vector Quantization), GMM-UBM (Gaussian Mixture Model-Universal Background Model), GSV-SVM (Gaussian Mean Supervector-Support Vector Machine), JFA (Joint Factor Analysis), IVEC ( authentication vector) and so on. Among them, the GSV-SVM system is superior to other systems due to its flexibility and robustness, and is currently wid...

AI Technical Summary

Problems solved by technology

[0005] In the current GSV-SVM speaker recognition system, often due to the short duration of the speech segment of the training speaker or the short duration of the test speech segment, some GMM mixing degrees cannot be obtained in the process of adaptive GMM mean supervector Fully adaptive, it is easy to cause the problem of "model hole", which limits the performance of speaker recognition in short speech

Images