摘要
传统的基于统计的子空间学习算法如主成分分析,通过学习只能得到一系列特征脸,忽略了人脸识别中重要的局部信息(如眼睛、鼻子)。而利用到类别信息的算法如线性判别分析,也会因为小样本问题而有所影响。为了解决这些问题,结合二维偏最小二乘与非负矩阵分解的非负性思想提出二维非负偏最小二乘(Two-Dimensional Nonnegative Partial Least Squares,2DNPLS)算法。其核心思想是在提取人脸特征时加入了非负性约束,使得2DNPLS不仅拥有偏最小二乘算法加入类别信息带来的分类效果,还保留了图像矩阵的内部结构信息,而且还使得到的基矩阵具有非负的局部的可解释性。在ORL,Yale人脸库中的实验结果表明,该算法从时间上和识别率上均优于人脸识别的主流算法。
Traditional subspace statistic methods, such as Principal Component Analysis （PCA） can only get a series of eigen face through learning, the available local features （eyes, nose） for face recognition are ignored. However, these methods incorpo- rating the category information such as Linear Discriminant Analysis （LDA）, face small sample problems. In order to take over these disadvantages, the paper proposes a novel approach to Extract the facial features called Two-Dimension Nonnegative Partial Least Squares （2DNPLS）. The main idea of the approach is grabbing the local features via adding the constraint of nonnegative to 2DPLS, which makes the approach gain not only the advantages of 2DPLS, incorporating both inherent structure and category information of images, but also the local features, having nonnegative interpretability. For evaluating the approach＇ s performance, a series of experiments are conducted on two famous face image databases ORL, Yale face databases, which demonstrate that the proposed approach outperforms the state-of-art algorithms.
出处
《计算机工程与应用》
CSCD
2013年第20期193-197,221,共6页
Computer Engineering and Applications
基金
国家自然科学基金(No.61173131)
中央高校基金(No.CDJZR12090002
No.CDJXS11100046
No.CDJXS11181162)
关键词
二维偏最小二乘
非负性
人脸识别
二维非负偏最小二乘
Two Dimension Partial Least Squares （2DPLS）
nonnegative
face recognition
Two Dimension Nonnegative PartialLeast Squares（2DNPLS ）
