基于肤色和Adaboost算法的人脸检测.pdf

基于肤色和Adaboost算法的人脸检测
摘要:人脸检测是指判断图像中是否有人脸,如果有人脸,则确定人脸在图像中的
位置和大小。人脸检测起源于人脸识别课题的研究,进入90年代,人脸检测开始作
为一个独立的课题受到研究者的重视。随着生物技术的的快速发展,人脸检测技
术在人脸识别、表情识别、视频会议、基于内容的索引、图像与视频检索、智能人
机交互等方面获得了广泛的应用。基于AdabooSt算法的人脸检测方法,具有检测
速度快,正确率高的优点,但是其误检率也比较高,所以需要在检测出人脸候选区
域后进行验证,将非人脸迅速排除掉。由于这种算法没有考虑到肤色因素,因此本
文提出在使用基于Adab。。St算法的人脸检测方法检测出人脸后,添加一个使用肤
色检测进行验证的过程。即在检测出人脸候选区域后,通过在YCbCr色彩空间中
建立的肤色模型对候选区域作进一步的肤色验证,排除掉候选区域中不是肤色的
区域,即排除掉那些被误检为人脸的区域。
关键字:人脸检测,Adaboost算法,肤色验证,色彩空间
Abstract:The definition of face detection is that determine whether
or not there are any faces in the image and,if Present,return the
image loeation and extent of eaeh detection originated in
face recognition on the subjcet,Since the 1990s,face detection began
as an independent subjcet of attention by the fast
rapid development of biological technology,face detection technology
inface
recognition,expression recognition,video conferencing,eonient based
indexing,image and video search,intelligent puter
interaction,obtain the widespread algorithm
which has high detecting speed and hit rate,bu t it also has high miss
rate,so it is neeessary to verify the face region safer detecting the face
and date regions and removes the none face regions quickly .Since
the algorithm has not taken the skin color information into
account,this paper proposes that it increases askin color detection to
verify the Face and date regions which was detected by using the
mehtod of based on Adaboost after detecting the
face candidate regions,it uses the skin color model which established
inthe YCbCr colors pace to make the further skin color
confirmationand removes in the candidate regions which has not the
skin color information,namely rules out the noneface region which
has detected by mistake.
Keywords: face detection, adaboost algorithm, skin color verification,
color space
1. 研究背景与目的

长期以来,计算机就好像一个盲人,需要被动地接受由键盘、文件输入的信息,而不
能主动从