直方图在线更新的肤色识别方法研究  

Research of skin color identification based on online update histogram

作　　者：杨扬[1] 程楠楠[1] 张萌萌[1] 

机构地区：[1]北方工业大学信息工程学院,北京100144

出　　处：《计算机工程与应用》2012年第4期204-206,共3页Computer Engineering and Applications

基　　金：国家自然科学基金(No.60013113)

摘　　要：提出了一种基于直方图在线更新的肤色识别方法,在使用Adaboost算法提取出人的脸部区域后,计算出该区域YCrCb色彩空间的颜色直方图,构建肤色区域的概率模型。使用该模型对图像中所有的像素点进行计算区分,并根据实时识别结果在线更新,达到较好的识别率。It proposes a skin color identification method based on the online histogram updating, the human face region is extracted by using Adaboost algorithm, and the region skin color histogram is calculated in YCrCb color space, and the probability model of skin color region is constructed. The model is used to calculate and differentiate all the pixels in the image and update the model according to real time result to achieve better recognition rate.

关 键 词：直方图 肤色识别 在线更新 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]