人脸识别中去光照算法  

Algorithm of Illumination Invariant in Face Detection and Recognition

作　　者：谭玉波[1] 田歌[1] 

机构地区：[1]河南工业大学,河南郑州450001

出　　处：《信息工程大学学报》2014年第5期603-608,共6页Journal of Information Engineering University

摘　　要：光照变化问题是人脸检测和识别中的一个非常重要而又难于解决的问题。提出一种人脸光照无关图算法,从单一图像中获得人脸光照的无关灰度图。算法在色度比空间中计算出信息熵最小的灰度图,所获得的灰度图像去除了光照影响,并且保留了图像信息和人脸特征。The effect of illumination changing is a very important issue and a problem in face detection and recognition. This paper presents an algorithm of a face illumination invariant,and the independent grayscale of face illumination is obtained from a single image. The algorithm calculates the minimum gray information entropy in color space. Gray image obtained by the algorithm removes the effect of illumination,while retains the image information and facial features.

关 键 词：去光照 信息熵 光照无关图 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]