基于感兴趣区域及LOD模型的图像压缩算法研究  

The Study of Image Compression Algorithm Based on ROI and LOD Model

作　　者：王冬[1] 杜扬[1] 孙鑫 李康宁[1] 

机构地区：[1]后勤工程学院军事供油工程系,重庆400016 [2]中国航空油料集团公司,北京100088

出　　处：《后勤工程学院学报》2007年第3期58-61,共4页Journal of Logistical Engineering University

基　　金：重庆市应用基础研究资助项目(2004BB2141)

摘　　要：感兴趣区域ROI（Region of Interesting）模型和层次细节LOD（Level of detail）模型是静止图像压缩及网络传输技术中两个重要模型。采用基于认知心理学的图像感兴趣区域提取模型，通过显著图法自动地得到与人眼视觉较吻合的ROI区域，并使ROI区域获得比背景区域更好的图像效果。同时通过将图像网格化，并在图像网格上进行B样条细分小波变换．实现了图像在不同分辨率下LOD压缩算法。ROI（Region of Interesting）and LOD（ Level of Detail ） model are two important techniques for still image compression and network transmission. In this article, the ROI retrieval model based on cognitive psychology is adopted. The regions of interesting extracted through saliency map method are almost the same with HVS （ Human Visual System）. Then the re- gions of interesting can get better visual effects than background region. The LOD scheme is used to realize the image compression at different resolution by image meshing and B spline subdivision wavelet transform.

关 键 词：ROI模型 LOD模型 图像压缩 

分 类 号：TN911.73[电子电信—通信与信息系统]