Emotion recognition method and device, electronic equipment and storage medium

Abstract

The invention discloses an emotion recognition method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring a to-be-processed target video file, wherein the target video file comprises an image and voice of a target user; extracting multiple groups of image sets and voice segments from the target video; analyzing the image set to obtain a face analysis result of the target user, and analyzing the voice segment to obtain an audio analysis result and a semantic analysis result of the target user; and determining a target emotion type of the target user according to the face analysis result, the semantic analysis result and the audio analysis result. The image set and the voice segment are extracted from the video file, and the facial analysis result, the semantic analysis result and the audio analysis of the user can be determined based on the image set and the voice segment, so that the emotion change of the user is positioned in multiple dimensions, and compared with the existing emotion recognition purely based on the user text content, the emotion recognition efficiency is improved. And the accuracy of identifying the emotion change of the user is greatly improved.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and in particular to an emotion recognition method, device, electronic equipment and storage medium. Background technique [0002] With the development of machine learning, artificial intelligence, and human behavior research, people pay more and more attention to the research of human emotional expression. Traditional technology for emotion recognition generally simply uses voice or vision to identify people's current emotions. This single recognition method often has a low accuracy rate and misidentification occurs. And relying solely on visual methods to recognize people's emotions, some people show different facial expressions when they are in joy, anger, sadness and other emotions; or only use voice to recognize people's emotions, usually the recognition results will be Affected by the dialect, whether the speaker's teeth are clear, and the surrounding environment,...

AI Technical Summary

Problems solved by technology

Traditional technology for emotion recognition generally simply uses voice or vision to identify people's current emotions. This single recognition method often has a low accuracy rate and misidentification occurs.

And relying solely on visual methods to recognize people's emotions, some people show different facial expressions when they are in joy, anger, sadness and other emotions; or only use voice to recognize people's emotions, usually the recognition results will be Affected by the dialect, whether the speaker's teeth are clear, and the surrounding environment, the accuracy of speech detection is reduced

Eventually lead to the inability to accurately identify people's emotions

Images