四叉树预测及其在图象压缩中的应用

QUADTREE PREDICTION AND ITS APPLICATION FOR IMAZE COMPRESSION

本文简述了四叉树及其预测原理,并提出了一种新的预测编码算法。它建立在四叉树图象描述方法的基础上,将以数组形式表示的数字图象转变为预测四叉树,并分层记录、处理及编码传输。该算法引入了自适应技术,具有较大的灵活性,可适当调整参数以满足对于图象质量和压缩比的不同要求。基于这种算法的数据压缩可以在码率为0.453b/p的条件下产生人眼视觉可接受的图象(峰峰信噪比>30dB)。

This paper briefly introduces quadtree data structure and its principle for the prediction, and presents a new predictive coding scheme. Basing on the quadtree image representation, a digital image which is recorded in matrix is converted to a predictive quadtree firstly, then is hierarchically recorded, processed and coded. Since using adaptive techniques with more flexibility, the algorithm can properly adjust some parameters to meet the different requirements of image quality and compression rate. This data compression technique can produce acceptabee images (PSNR> 30dB) at rate of 0.453b/p.