Voiceprint recognition model training method and device, equipment and readable medium

Abstract

The invention discloses a voiceprint recognition model training method, apparatus and device, and a readable medium. The voiceprint recognition model training method comprises the following steps: obtaining an effective voice frame of training set voice; extracting the MFCC features of the effective voice frames, and performing difference expansion processing on the MFCC features of the adjacent effective voice frames; and inputting the MFCC features after the difference expansion processing into an initial voiceprint recognition model, training the initial voiceprint recognition model until convergence, and obtaining a trained voiceprint recognition model. According to the voiceprint recognition model training method provided by the embodiment of the invention, effective voice analysis is performed on the training set voice, the MFCC features of the effective voice frames are extracted, and difference expansion processing is performed on the MFCC features of the adjacent effective voice frames, so that the change trend of the adjacent effective voice frames is highlighted, the effectiveness of voiceprint feature extraction can be effectively improved, and the voiceprint recognition efficiency is improved. Therefore, the voiceprint recognition model with high voiceprint recognition accuracy is obtained.

Description

technical field [0001] The present application relates to the technical field of voiceprint recognition, in particular to a voiceprint recognition model training method, device, equipment and readable medium. Background technique [0002] Voiceprint Recognition (Voiceprint Recognize) is a technology that automatically identifies the speaker's identity based on the voice parameters in the voice waveform that reflect the speaker's physiological and behavioral characteristics. With the wide application of voiceprint recognition technology in public security, financial anti-fraud, criminal investigation, etc., there are more and more application scenarios and the requirements for related technical indicators are getting higher and higher. [0003] The existing voiceprint recognition scheme is to collect voice data of at least one user in advance, extract feature values ​​from the voice data, and input the extracted feature values ​​into a voiceprint model to obtain an N-dimensio...

AI Technical Summary

Problems solved by technology

However, in the existing technology, the background environmental noise of the voice to be tested has a great influence on the accuracy of voiceprint recognition, which in turn leads to the low accuracy of the existing voiceprint recognition.

Images