报告标题: Face Recognition Algorithms based on orthogonal and sparse constrained nonnegative tensor factorization

报告人：徐常青 （苏州科技大学教授）

报告摘要:  We propose a face recognition algorithm based on orthogonal and sparse constrained nonnegative tensor factorization, which improves the accuracy of the traditional face recognition approaches. We first add the orthogonal and sparse constraints to the traditional nonnegative tensor factorization to reduce the correlation between the base images and obtain sparse coding. Then we use the original face image and the decomposed base image to calculate the low dimensional feature representation of the face. The cosine similarity is used to measure the similarity between low-dimensional features and judge whether two face images represent the same person. The experimental results show that our algorithm can achieve better recognition results with an improved accuracy. （This is a joint work with my graduate student Ms. Shan Song. ）

报告时间：2020年11月12日 (周四)上午9:00-10:00

报告地点：腾讯会议（会议ID：896885729）

报告人简介：徐常青，博士，苏州科技大学教授、苏州市高校科研院所紧缺高层次人才、中国教育数学专委会常务理事、江苏省运筹学会理事、上海大学“张量与矩阵计算国际研究中心”学术委员会成员, 美国数学评论《Mathematics Review》评论员；曾任浙江农林大学数学系主任兼应用数学学科负责人；Linear Algebra Applications、Linear Multilinear Algebra、Electronic Journal of Linear Algebr等杂志审稿人。在SIAM, Linear Algebra Applications、Linear Multilinear Algebra等SCI期刊上发表论文90余篇，主编教材3部。曾访问美国Univ of Pacific、San Jose State Univ和斯坦福大学（短期），2013-2017年多次应邀访问香港理工大学。主要研究工作为非负张量分解、完全正张量、范德蒙张量及其在统计学方面的应用等。申报发明和实用性专利5项，其中3项已获批。

邀请人：喻高航