一种基于感兴趣区域的图像近无损压缩方法  

A Near-lossless Image Compression Algorithm Based on Regions of Interest

作　　者：魏琳[1] 马大玮[1] 余训锋[1] 

机构地区：[1]重庆大学通信工程学院测控研究所,400044

出　　处：《中国科技信息》2006年第24期80-81,83,共3页China Science and Technology Information

基　　金：重庆市自然科学基金(编号:CSTC2005BB2206)

摘　　要：感兴趣区域(ROI)图像近无损压缩方法,即在ROI采用无损压缩,在背景区域采用有损压缩,从而保证重要信息不丢失,并且尽可能提高图像的压缩比。本文算法对ROI使用Huffman编码,背景区域使用EZW编码,并且为了增大压缩比,对EZW编码码流再进行算术编码。实验结果表明,随着ROI面积的增大,图像的压缩比减小,信噪比增加。在ROI区域占整幅图像大小10%以下时,该方法的实用性较好,能得到较大的压缩比和较好的重构图像。Near-lossless image compression algorithm based on Regions of Interest （ROI） means to compress interesting regions of an image Iossiess, and to compress background regions With a little loss,Thus guarante the Important information fidelity,and to increase image compressing ratio as far as possible.This paper brings forward an algorithm which realizes neariossiess image compression by using the Huffman encoding in ROI and the EZW In background region, In order to increase the comPreSsion ratio, carries on the arithmetic code again to the EZW code symbol stream. The experimental results indicates image compression ratio decreased and the signal-to-noise ratio incresod as the size of ROI increasing. When the ROI region accounts for the entire picture size below 10%,this method usability is better,can obtain an greater compression ratio and the better composition picture.

关 键 词：感兴趣区域(ROI) 整数小波变换 EZW 图像压缩 

分 类 号：TN919.8[电子电信—通信与信息系统]