一种基于光照方向估计的人脸识别方法

Face Recognition Based on Illuminant Direction Estimation

针对不同光照条件下的人脸识别率较低,尤其是在极端光照条件下识别率急剧下降的问题,提出了一种基于垂直积分投影和高光区域处理的光照方向估计方法。利用该方法,可以根据估计的光照方向对不同的光照条件下的图像进行分类。经过分类后,将每个光照类别对应一个投影子空间,然后将分好类的图像分别投影到各自的子空间进行识别,以提高识别效果。最后在YaleB人脸数据库上进行实验,实验结果表明该方法可有效地提高在不同光照条件下的人脸识别率。

To solve the problem that face detection rates usually drop quickly if the illumination is too bright,too dark,or non-uniform illumination ,present a method based on estimated illumination and vertical integral projection. First of all, the images are classified according to the illumination conditions, and then protected to each subspace to be recognized,improving the recognition effect. The experimental results on YaleB face database show that this method can effectively improve the face detection rates under different illumination conditions.