Deep Learning Emotion Recognition Method

Abstract

Emotion recognition refers to the process of actively analyzing human emotions through computer technology, and it has become an important part of modern society. Traditional emotion recognition is mainly based on a single information source, such as text, speech, video, etc., from which emotional features are extracted for classification or regression to recognize human emotions. With the development of artificial intelligence technology, multimodal emotion recognition is gradually becoming widely used. It combines two or more types of information, such as text, speech, and visual information, in different ways to analyze emotions. Multimodal emotion recognition is far superior to a single modality in understanding emotions. This article mainly analyzes the technology of emotion analysis. Firstly, we introduce the basic concepts and research status of emotion recognition. Then, we introduce the main types of emotion recognition and describe various methods used in the process in detail. Finally, we discuss the challenges and future developments of emotion recognition.