摘要
随着科学技术的进步和发展,图像已经成为了信息的一个重要来源,而图像中的重要信息往往只集中在部分区域。而且图像数据在信道中传输时,可能会发生数据包丢失或出错的问题。鉴于以上问题,在对图像ROI和多描述量化编码进行分析和综合的基础上,本文提出了一种结合图像感兴趣区域(ROI)提取和多描述量化,零树编码的图像传输方法。根据图像特征选定ROI区域,然后对ROI进行多描述量化编码,对其他区域进行普通的量化编码,试验表明此方法可以得到更好的重构图像。
Along with the science technical progress and the development, the image has already become an important source of information, but the more important information usually locates in some parts. While delivered, data pack usually lose due to channal station.Considering this all, this text putforward a image Iransmitting algorithm combining the regions of interest (ROI) and multiple description Quanfizafion and EZW code, with realize their character. Frist select ROI according to the picture, then quanfilize the ROI with multiple description quantizer, the other parts with normal Quantization. Experimentation testifys this "algorithm is better than others.
出处
《自动化与仪器仪表》
2009年第3期81-82,119,共3页
Automation & Instrumentation
关键词
感兴趣区域(ROI)
多描述量化
零树编码
Regions of Interest(ROD Multiple description quantization
Zero-Tree Coding(EZW)