基于小波变换的ROI图像压缩改进算法研究  

A Study of the Improved Algorithm of ROI Image Compression Based on Wavelet Transform

作　　者：王晓芳[1] 邱书波[1] 张绪光[1] 

机构地区：[1]山东轻工业学院电子信息与控制工程学院,山东济南250353

出　　处：《计算机工程与科学》2010年第8期84-86,133,共4页Computer Engineering & Science

基　　金：山东省自然科学基金资助项目(Y2008G15)

摘　　要：小波变换作为一种多分辨率分析方法,具有良好的空频局部特性,特别适合按照人类视觉系统特性设计图像压缩编码方案。基于小波变换的嵌入式编码算法能够有效地应用于基于感兴趣区域(ROI)的图像压缩与渐进传输。为了进一步提高效率并改善图像压缩传输的效果,本文在该算法的基础上提出了两点改进:(1)计算感兴趣区域模板时,将数学上的"排列组合"思想应用于传统的回溯法,使计算过程更加简单;(2)在对小波系数嵌入式编码时,牺牲图像背景区域的细节系数以获取感兴趣区域更高的解码质量。最后给出了改进算法的部分实验结果,证明了改进算法的有效性。Wavelet transform, as a method of multi-resolution analysis, which has good characteristics of space-frequen- cy , is especially suitable for image coding based on human vision. The algorithm of embedded coding based on wavelet transform can be applied to image compression and progressive transmission effectively. In this paper, the traditional algorithm is improved in two ways to increase the efficiency of image coding and transmission. First, the idea of permutation and combi- nation in mathematics is applied to the traditional method of backtracking in calculating the ROI model. Second, the wavelet coefficients of the background are rejected from the embedded coding stream of wavelet coefficients to achieve higher quality of decoded ROI images. Some experimental results of the improved algorithm are provided at the end of this paper to prove the validity of the improved algorithm.

关 键 词：图像压缩 感兴趣区域 排列组合 嵌入式编码 小波变换 

分 类 号：TN919.81[电子电信—通信与信息系统]