摘要
提出了一种小波域内相邻子图的分形编码预测方法。首先选择合适小波基对图像进行二级分解,然后对三个二级高频子图进行Jacqain方法的分形编码,利用小波域内同方向不同分辨率子图之间的相似性,对得到的分形编码参数进行适当比例变换,最终从水平、垂直和斜方向等三个方向近似预测出对应的一级高频子图分形编码。实验表明,与Jacqain相比本方法在同类图像基本保证解压图像质量的同时,极大地缩短了编码时间和提高了压缩比。
A fractal coding prediction method based on wavelet domain is put forward. Firstly based on suitable wavelet base, a pyramid subband decomposition with two levels is done. Then three high frequency subbands of two-level are coded independently by Jacqain method. According to the similarity in structure between different subbands with same orientation,with suitable linear scale thransform,fractal codes in three high frequency subbands of one-level throughout horizontal,vertical and oblique direction are predicted approximately. The experimental results show that the method shortens coding time and lifts compression ratio,which is superior to Jacqain fractal block coder.
出处
《计算机应用与软件》
CSCD
北大核心
2007年第6期68-70,共3页
Computer Applications and Software
基金
浙江省教育厅科研项目(编号:20060183)
关键词
小波域
分形编码
预测
编码时间
Wavelet domain Fractal coding Forecasting Coding time