Multi-mode-based conference spokesman identity non-inductive confirmation method

Abstract

The invention provides a multi-mode-based conference speaker identity non-inductive confirmation method. Based on a conference using multiple modes of image, voice and text, the identity of a spokesman is confirmed by recognizing the expression, voice and speaking style of the spokesman, and the method specifically comprises an expression recognition method based on a deep learning model, a voicerecognition method based on an artificial intelligence algorithm and a method for recognizing speaking content by adopting a text clustering algorithm. According to the method, the whole process is automatic, manual intervention is not needed, the identity of the speaker can be confirmed in a non-inductive mode through the artificial intelligence algorithm model, manual intervention is not needed,meeting and office efficiency is greatly improved, and accuracy is high.

Description

technical field [0001] The invention belongs to the field of natural language processing, in particular to a method for non-sensing confirmation of the identity of a conference speaker based on multimodality. Background technique [0002] With the development of the economy, efficient office is increasingly inseparable from the conference system. At this stage, many conference systems need to record the speech content of each speaker for the convenience of summarization and reporting. Therefore, for this requirement, an intelligent and fast method for distinguishing speakers is needed. [0003] At present, the current conference system mostly uses the microphone to record the voice of the speaker to record the content of the speech. If you want to distinguish different speakers, you need to assign a microphone to each speaker. However, if you assign multiple microphones, it may cause crosstalk. Because the distance is too close, multiple microphones will recognize a person ...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the cumbersome problem of distinguishing different speakers due to the need to turn off and turn on the microphones at different positions many times due to the adjustment of the distance caused by the traditional allocation of microphones in regular meetings, the present invention provides a multi-modal The method for confirming the identity of the conference speaker without feeling, specifically: the method of automatically identifying and distinguishing the speaker's expression, voice and speech style in three aspects, including the expression recognition method based on the deep learning model, the method based on Voice recognition method based on artificial intelligence algorithm, speech content recognition method based on text clustering algorithm

Images