基于感兴趣区域的小波图像压缩算法  

A method of wavelet image coding based on region-of-interest

作　　者：乔慧[1] 郭强[2] 张书玲[1] 

机构地区：[1]西北大学数学系,陕西西安710069 [2]西安科技大学基础课部,陕西西安710054

出　　处：《纺织高校基础科学学报》2004年第2期150-154,共5页Basic Sciences Journal of Textile Universities

基　　金：陕西省自然科学基金资助项目(2001SL03)

摘　　要：提出了一种基于矢量量化的感兴趣区小波图像压缩算法,该算法充分利用了子带系数带间和带内的相关性并考虑到人眼对背景区和感兴趣区的视觉差异,对背景区和感兴趣区分别采用不同的矢量构成方法,且在对背景区编码时借用零树编码的思想极大的提高了编码比特率.在高压缩比时应用本算法所得恢复图像仍能保持较好的主观质量.Based on region of interest(ROI) ,a wavelet image coding method using vector quantization(VQ) is presented .This algorithm exploits the correlation between the subbands and within the subbands fully.According to the different priority of ROI and region of background(BG),it uses different method to combinate vector of them.Especially,the utilizing of zerotree coding's idear in the BG coding significantly increase the bite rate.With human characteristics to best advantage,reconstructed image with a high compressive rate can still hold fairly good subjective quality.

关 键 词：小波编码 矢量量化 感兴趣区 

分 类 号：TN911.73[电子电信—通信与信息系统]