摘要
针对人脸光照、遮挡、身份、表情等因素变化的人脸姿态估计难题,结合稀疏表示分类(SRC)方法的优秀识别性能,对SRC理论进行了深入分析,并将其应用于人脸姿态分类。为了解决姿态估计中人脸光照、噪声和遮挡变化问题,将人脸姿态离散化为不同的子空间,每个子空间对应一个类别,据此,提出基于字典学习与稀疏约束的人脸姿态识别方法。通过在公开的XJTU和PIE人脸库上实验表明:所研究的方法对人脸光照、噪声和遮挡变化具有鲁棒性。
According to the challenges in face pose estimation under different illuminations, occlusions, identity, expressions, and so on, combining with the excellent classification performance of sparse representation classification （ SRC ）, a deep analysis on the theory of SRC and its application in face pose classification are made. In order to handle challenges such as variation of face illumination, noises and occlusion, a robust face pose estimation method based on dictionary learning and sparse representation is presented. In which face poses are discrete into different subspaces, each subspace corresponding to a class. Several experiments are performed on XJTU and PIE databases. Recognition results show that the proposed method is suitable for efficient face pose recognition under illumination, noises and occlusion variations.
出处
《电视技术》
北大核心
2015年第13期40-44,共5页
Video Engineering
基金
国家自然科学基金项目(61271256)
河南省重大科技攻关项目(072SGZS38042)
湖北科技学院博士启动基金项目(BK1418)
关键词
人脸姿态估计
稀疏表示
子空间学习
人脸识别
face pose estimation
sparse representation
subspace learning
face recognition
