一种基于小波变换的邻域搜索分形图像编码

A neighbor-region search fractal image coding based on wavelet transform

本文根据文献的研究结果,对传统的分形图像编码算法加以改进,即使用邻域搜索法代替传统的全局搜索法。并将其运用到图像小波变换后的低频部分,同时为得到较好的图像恢复效果,对高频部分采用进行零树编码。实验表明,该方法在较好的恢复图像的前提下,大大的提高了编码的速度。

In this paper, the neighbor-region search fractal image coding method is proposed to take the place of the traditional fractal image coding according to the research results of references 1 and 2. We apply the improved fractal image coding algorithm only to the low-frequency coefficients after the wavelet transform, and to get a better image recovering, zero-tree encoding algorithm is applied to the height-frequency coefficients. Experiment results show that the coding time is reduced evidently while the quality of reconstructed image is the same.