基于图像子空间的改进商图像方法

An Improved Quotient Image Method Based on Image Subspace

光线变化将显著降低人脸识别系统的性能。Shashua et al.提出了一种处理人脸识别中的光线变化问题的简便方法——商图像方法。在本文中,我们从图像子空间的角度对商图像方法进行了分析,理论分析和实验表明,这种方法存在的主要缺点有:1)不准确的理想类假设;2)简单的三维点光源模型无法很好地近似任意光照情况。针对这些不足,我们提出了一种基于图像PCA子空间的改进商图像方法,以克服这些缺点。我们的方法能够较好地满足商图像方法的理论前提,从而达到更好的图像合成效果和人脸识别性能。

Illumination changes can significantly decrease the performance of face recognition systems. In this paper,we present an analysis of quotient image method from the view of image linear subspace,and point out its main deficiencies： 1） inaccurate assumption of ideal class of object, 2） 3-d subspace doesn＇t suffix to represent all images of an object under varying illuminations. An improved method based on image subspace constructed using principal component analysis （PCA） is presented to overcome these deficiencies. Our method can meet the theoretical prerequisites of quotient image method,and achieve better image synthesis and face recognition results.