摘要
首先,针对人脸表情识别问题提出一种新的多尺度特征选择网络识别方法,该网络充分结合多尺度网络结构和特征选择结构的优点,能更有效地提取面部静态图像中的空间信息.其次,为验证本文提出的多尺度特征选择网络的识别性能和泛化能力,在两个经典的人脸表情识别数据集上与一些常用的方法进行对比和交叉验证实验.实验结果表明,该网络取得了更好的识别效果,并且具有良好的泛化能力,可以灵活地嵌入到人脸表情识别分析系统中.
Firstly,aiming at the problem of facial expression recognition,we proposed a new multi-scale feature selec tion network recognition method.The network fully combined the advantages of multi-scale network structure and feature selection struc ture,which could extract the spatial information in the facial static images more effectively.Secondly,in order to verify the recognition performance and generalization ability of the proposed multi-scale feature selection network,we carried out comparison and cross validation experiments with some common methods on two classical facial expression recognition databases.The experiment results show that the proposed netw ork achieves better recognition effect and has good generalization ability,it can be flexibly embedded into the facial expression recognition and analysis system.
作者
齐妙
闫光友
徐慧
孙慧
QI Miao;YAN Guangyou;XU Hui;SUN Hui(College of Information Science and Technology,Northeast Normal University,Changchun 130117,China;Institute of Technology,Changchun Humanities and Sciences College,Changchun 130117,China)
出处
《吉林大学学报(理学版)》
CAS
北大核心
2022年第2期425-431,共7页
Journal of Jilin University:Science Edition
基金
国家自然科学基金青年科学基金(批准号:61907007)
吉林省科技厅工业领域项目(批准号:20200401086GX,20200401081GX)
吉林省教育厅“十三五”科学技术研究项目(批准号:JJKH20190355KJ).
关键词
卷积神经网络
人脸表情识别
特征选择机制
多尺度网络
convolutional neural network
facial expression recognition
feature selection mechanism
multi-scale network