图像无损压缩的预测编码及量化误差处理  

Predictive Coding and Disposal of Quantized Error for Lossless Compression of Images

作　　者：王文成[1] 卢金明[1] 张永林[1] 

机构地区：[1]暨南大学物理系,广州510632

出　　处：《光电子．激光》2004年第5期608-611,共4页Journal of Optoelectronics·Laser

摘　　要：本文提出了一种基于预测编码并对量化误差进行四叉树编码从而实现对灰度图像无损压缩的方法。采用固定预测系数的算法,使得预测过程速度很快;通过四叉树编码方法处理量化误差,避免了对小数编码和处理这一复杂的过程。整个实现过程十分简单,实验结果显示,图像通过预测器后的平均码长明显小于原始图像的熵。In this paper we propose a new method for lossless compression of grayscale images based on predictive coding, in which quantized error is processed by quadtree coding. The predictive arithmetic makes the process of prediction quicker and easier by using fixed coefficient. And through quadtree coding method for quantized error, complicated course of decimal fraction is avoided. The whole process is easy to realize. The result of experimentation shows that the final issue of prediction is much smaller than the entropy of original image.

关 键 词：预测编码 四叉树 图像无损压缩 量化误差 

分 类 号：TN919.81[电子电信—通信与信息系统]