摘要
为解决不均匀的光照分布影响人脸身份识别的准确率问题,提出了一种无约束光照条件下的人脸身份识别方法.该方法首先基于改进的Retinex理论对人脸灰度图像进行光照预处理;然后对处理后人脸图像中鼻子周围的遮挡阴影区域进行光照补偿;最后利用主动外观模型定位人脸特征点,并选取关键特征点计算人脸身份特征,进而实现人脸身份识别.基于扩展Yale人脸数据库和CMU-PIE数据库进行实验,结果表明所提方法能有效改善人脸图像的光照情况,提高人脸身份识别率.
In order to solve the problem that uneven illumination distribution seriously affects the accuracy of face recognition,a face identity recognition method under unconstrained illumination was proposed.First,illumination preprocessing of grayscale face image was performed based on the improved Retinex theory.Then,illumination compensation was applied to the shadow region around the nose in the preprocessed face image.Finally,active appearance model was utilized to mark the facial feature points,and some key points were selected to calculate the face identity features.Experimental results using the extended Yale face database B and CMU-PIE database show that the proposed method can achieve better illumination normalization and improve the face identity recognition rate.
作者
易积政
阳洁琼
吴星宇
赵立明
YI Jizheng;YANG Jieqiong;WU Xingyu;ZHAO Liming(College of Computer and Information Engineering,Central South University of Forestry and Technology,Changsha 410004,China;College of Life Science and Technology,Central South University of Forestry and Technology,Changsha 410004,China;School of Medicine,University of Pennsylvania,Philadelphia 19104,USA;School of Advanced Manufacturing,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2020年第1期41-47,共7页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(61602528,51604056)
湖南省教育厅科学研究项目优秀青年资助项目(16B275)
中南林业科技大学高层次人才引进资助项目(2015YJ013)
国家留学基金委出国访学资助项目(201808430002)
湖南省自然科学基金资助项目(2017JJ3527)
关键词
人脸身份识别
无约束光照
光照均衡
人脸特征点
身份特征
face identity recognition
unconstrained illumination
illumination normalization
facial feature point
identity characteristics
