子模式局部保持映射人脸识别

摘要

研究表明基于整体思想的人脸识别方法由于忽略图像的局部信息,在识别性能方面不如局部信息特征保持较好的基于子模块思想的识别算法.基于应用流形技术对图像降维后能够较好保持非线性子流形中的局部数据流形结构,提出了一种改进的子模式局部保持映射人脸识别算法.其主要思想是将同类的不同图像一并划分子集,由同位置子图组成子模块,并对子模块运用LPP算法学习其流形结构,与将不同类图像一并划分子集学习流形的方法不同.实验表明,该算法能更好地保持人脸图像的局部流形结构和信息特征,提高了识别率.%Researches show that sub-pattern based face recognition approaches perform better than whole image based methods in local face information preservation. As manifold learning technologies preserve local manifold structure of the nonlinear sub-manifold while implementing dimension reduction, this paper put forward a sub-pattern locality preserving projection ( BspLPP). Unlike previous approaches partitioned all training images of different classes into sub-images and used the same location images to form a sub-pattern, BspLPP first partitioned the same class images into different sub-images, used the same location sub-images to form a sub-pattern, and then applied LPP to learn the manifold structure of each sub-pattern. Experimental results show that BspLPP preserves the manifold and local information well and improves the recognition performance.