分块NSA在人脸识别上的应用

Modular NSA's application to face recognition

基于非参数子空间分析(nonparametric subspace analysis,NSA)方法,提出了分块NSA方法并将应用于人脸识别上。分块NSA方法首先对图像矩阵进行分块,对分块得到的子图像矩阵再利用NSA进行鉴别分析。这样做有以下2个优点:1)能有效地抽取图像的局部特征,对人脸表情和光照条件变化较大的图像表现尤为突出;2)与NSA相比,由于使用子图像矩阵,分块NSA可以避免使用奇异值分解理论,过程简便。此外,NSA是分块NSA的特殊情况。在ORL和XM2VTS人脸库上验证了该方法在识别性能上优于NSA和分块LDA方法。

Modular NSA,is proposed which is applied to face recognition.Firstly,in the proposed approach,the original images are divided into modular images,which are also called sub-images.Then,the well-known NSA method can be directly employed to the sub-images obtained from the previous step.There are two advantages of the proposed method：1）local feature of the images can be extracted efficiently,and it is really true of the images that have large variations in facial expression and illumination;2）Singular value decomposition of matrix may be avoided in the process of feature extraction,which is simpler than that of other technologies such as NSA.Moreover,NSA is a special case of modular NSA.Experimantal results on ORL and XM2VTS face datafaces show the performance of modular NSA is obviously superior to NSA and Modular linear discriminant analysis.