基于JPEG2000渐进过程数截断编码算法  

Progressive Passing-number Truncation Coding Algorithm Based on JPEG2000

作　　者：孙文方[1] 赵亦工[1] 石自民[1] 宋蓓蓓[1] 

机构地区：[1]西安电子科技大学模式识别与智能控制研究所,西安710071

出　　处：《计算机工程》2006年第20期213-215,共3页Computer Engineering

基　　金："十五"国防预研项目

摘　　要：对JPEG2000的率失真优化截取的嵌入式码块编码算法的各编码块截断过程数进行了分析,对大量的图像进行实验后发现,图像的每个编码块是在相近的过程数截断,于是提出了一种同一过程截断算法。在此基础上又提出了渐进过程数截断编码算法,最终这个算法舍去了复杂的率失真计算,降低了T2编码器设计难度,提高了JPEG2000系统的处理速度,减少了存储量,对硬件实现是很好的新方法,处理512×512的8bits位深的图像16倍压缩为250fps。Some analysis is focused on each code block passing number of EBCOT （embedded block coding with optimized truncation） in JPEG2000. Through experiments on lots of images, a conclusion can be made that each code block truncates at similar code block passing numbers. Hence this paper addresses a truncation algorithm in the equal-coding pass. Furthermore based on the above algorithm, a progressive passing-number truncation coding algorithm （PPNTC） is proposed which discards complicated rate-distortion calculation and makes it easy for the design of T2 coder. This algorithm speeds up the processing of JPEG2000 and saves the memory cost. It can be realized more easily on hardware system. The system shows the developed scheme meets the high-speed requirement of the real-time coding 250fps.

关 键 词：JPEG2000 EBCOT 率失真优化 等编码过程 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]