基于RGB色彩模型相关模板的彩色人头图像编码  

COLOUR HEAD IMAGE ENCODING BASED ON CORRELATION TEMPLATE OF RGB COLOUR MODEL

作　　者：陈宇拓[1] 李建红[1] 杨炫[1] 韩旭里[2] 余英林[3] 

机构地区：[1]中南林业科技大学计算机科学学院,湖南长沙410004 [2]中南大学数学科学与计算技术学院,湖南长沙410083 [3]华南理工大学电子与通信工程系,广东广州510641

出　　处：《计算机应用与软件》2010年第5期40-44,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(60372068)

摘　　要：基于彩色人头图像RGB色彩模型色彩分量之间的相关性,提出一种针对彩色人头图像压缩编码的方法。该方法将彩色人头图像的面部特征区域用一种快速且较精确的方法提取出来,对色彩分量G面部特征区域进行细致编码,而对面部特征以外的区域进行粗犷编码,结合在色彩分量R、B与G之间建立的相关模板,实现对彩色人头图像的分级编码。实验结果表明该方法既能获得较满意的重构图像效果,又可大大提高编解码的效率和压缩比。This paper presents a method specifically for the colour head image compression encoding based on the correlation among three colour components of RGB colour model of the image.This method extracts face feature region of the colour head image with quick and fairly precise means.The face feature region of the colour component G is encoded elaborately,and the region not related to the face feature is encoded roughly.In combination of correlate template established among three colour components of R,B and G,the classified encoding of the colour head image is achieved.Experimental results show that this method can obtain satisfactory effect of the reconstructed image and can also improve the encoding/decoding efficiency and the compression ratio.

关 键 词：特征区域提取 相关系数 相关模板 分级编码 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]