Voiceprint clustering method, electronic equipment and storage medium

Abstract

The invention discloses a voiceprint clustering method, electronic equipment and a storage medium, and the voiceprint clustering method comprises the steps: obtaining face detection information synchronized with audio information; determining people number information based on the face detection information; and assisting the audio information to perform voiceprint clustering based on the number of people information. According to the method provided by the embodiment of the invention, the performance and the effect of a speaker log algorithm can be improved by combining the face information in the audio information and the video information, the speaker log is a preposed link of numerous voice tasks and can be suitable for subsequent voice recognition tasks, TTS and other voice-related tasks, and in addition, semantic understanding can be carried out, so that the user experience is improved. Other man-machine conversations are assisted, semantic results can form a knowledge graph, and the method has a wide application prospect.

Description

technical field [0001] The invention belongs to the technical field of voiceprint clustering, and in particular relates to a voiceprint clustering method, system, electronic equipment and storage medium. Background technique [0002] In related technologies, voiceprint verification is used to determine whether a speaker matches a known speaker. Voiceprint recognition, judging whether the speaker belongs to one of the multiple candidates in the database, that is, one of the multiple candidates is matched. Personalized speech detection, focusing on the content of a specific speaker. Single-modal speaker log, speaker log generally refers to the algorithm of single voice mode, that is, to record the log of the speaker from the perspective of pure audio. [0003] The inventor found that: the defect of voiceprint verification and voiceprint recognition is that it is only used to verify whether a certain speaker has a specific speaker, and does not record the moment when the spee...

AI Technical Summary

Problems solved by technology

[0003] The inventor found that: the defect of voiceprint verification and voiceprint recognition is that it is only used to verify whether a certain segment of speakers has a specific speaker, and does not record the moment when the speech occurs; the defect of personalized voice detection is that it only focuses on a specific speaker. The speaker's content and time of occurrence, rather than recording all the people who appear in the conversation; the defect of the single-mode speaker log is that it cannot work effectively in the case of high noise and overlapping human voices, and only borrows the voice mode information, not making good use of visual information

Images