【摘要】 本文提出一种基于卷积的快速分形编码算法。在求取分形压缩仿射变换参数的过程中 ,利用快速卷积来计算区域块与排列块之间的互相关 ,从而实现了编码过程的无损加速。为了在加速编码过程的同时提高压缩比 ,将快速卷积算法与基于四叉树分割的邻域搜索算法相结合。通过计算机仿真实验 ,获得了良好的效果。与经典的 PIFS方法相比 ,在信噪比大致相当的前提下 ,编码速度和压缩比都得到了显著提高。更多还原

【Abstract】 In this paper,a fast fractal coding algorithm based on convolution theory is presented.During the procedure of affine transform parameters calculation in fractal compression,fast convolution theory is used to calculate the cross- correlation between domain blocks and range blocks,in this way,the process of encoding is accele- rated losslessly.In order to raise compress ratio while accelerating encoding process,an algorithm com- bining fast convolution with quadtree partitioning based on neighbor search is adopted.Computer simulation has shown that under the condition of moderate loss of image quality,this algorithm brings up higher compression ratio and encoding speed compared with that of conventional PIFS method.更多还原