基于深度学习的情感识别研究综述  

Review of Emotion Recognition Based on Deep Learning

作　　者：张艳倪 汤敏[1] 王坤赤[1] Zhang Yanni;Tang Min;Wang Kunchi(School of Information Science and Technology,Nantong University,Nantong,Jiangsu 226019,China)

机构地区：[1]南通大学信息科学技术学院,江苏南通226019

出　　处：《计算机时代》2025年第3期1-5,共5页Computer Era

基　　金：国家自然科学基金项目“大深度饱和潜水智能氦语音通信系统研究”(No.62371261)。

摘　　要：深度学习是基于深层神经网络的机器学习,在情感识别技术领域已取得了显著的进步。本文综述深度学习在情感识别上的应用,分析其技术框架与研究现状;通过研究常用数据库特点、语音特征提取和情感分类技术的应用,进一步讨论基于深度学习的情感识别模型的评价指标体系,并对相关文献进行分类综述。研究表明,基于深度学习的情感识别准确率已达80%以上,应用于现实生活可期,本文对未来研究方向进行了展望。Deep learning is a type of machine learning based on deep neural networks,and it has made remarkable progress in the field of emotion recognition technology.The application of deep learning in emotion recognition is reviewed in this paper,the framework of deep learning is introduced,the current situation of emotion recognition research is analyzed,including the characteristics of the database,speech feature extraction and emotion classification.Then,the relevant literature is reviewed and the evaluation indexes of emotion recognition are introduced.The accuracy rate of emotion recognition based on deep learning has exceeded 80%,and it is possible to be applied in real life after further improvement and optimization in the future,the future development direction of emotion recognition based on deep learning is looked forward to.

关 键 词：情感识别 语音特征提取 深度学习 神经网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]