摘要
人脸表情识别就是让计算机按照人类的思维理解表情,是人机交互的重要组成,然而随着深度学习的迅速发展,深度学习技术在人脸表情领域的研究也成为研究热点,因此对深度学习技术在表情识别中的应用及取得的成果进行分析是有必要的。首先总结了几种常用表情数据集;然后从特征提取和特征分类两方面对基于深度学习的表情识别方法进行了分类,并从网络改进方面分析了基于深度学习的表情识别中的几种网络改进方法;最后阐述了表情识别这一领域中面临的挑战和未来发展。
Facial expression recognition makes computer understand facial expression according to human thinking,which is an important part of human-computer interaction.However,with the rapid development of deep learning,the research of deep learning technology in facial expression field has become a research hotspot,so it is necessary to analyze the application and achievements of deep learning technology in facial expression recognition.Firstly,several common expression data sets were summarized.Then,the expression recognition methods based on deep learning were classified from two aspects of feature extraction and feature classification,and several network improvement methods in expression recognition based on deep learning were analyzed from the aspect of network improvement.Finally,the challenges and future development in the field of expression recognition were described.
作者
党宏社
王淼
张选德
DANG Hong-she;WANG Miao;ZHANG Xuan-de(School of Electrical and Control Engineering,Shaanxi University of Science and Technology,Xi’an 710021,China)
出处
《科学技术与工程》
北大核心
2020年第24期9724-9732,共9页
Science Technology and Engineering
基金
国家自然科学基金(61871206)。
关键词
深度学习
表情识别
特征提取
表情分类
deep learning
expression recognition
feature extraction
expression classification
