| 摘 要: 为了解决分形图像编码耗时过长的问题，该论文主要研究了基于K-均值聚类的快速分形编码算法。首先 引入方差法将子块分为简单块和复杂块，随后采用K-均值聚类算法对复杂子块及父块进行分类，并在搜索匹配父块的 过程中运用近邻搜索法，使得相应子块仅在近邻范围内与同类的父块进行匹配运算。该方法对匹配块的搜索过程进行了 优化，大幅度减少了编码时间。测试结果表明，与基本分形编码算法相比可提速多倍，并且其重构图像效果较好。 |
| 关键词: 分形图像编码；K-均值聚类；近邻搜索；方差法 |
|
中图分类号: TP391
文献标识码: A
|
| 基金项目: 国家自然科学基金资助项目(61502343)；广西自然科学基金资助项目(2015GXNSFAA139295). |
|
| An Improved Fractal Coding Method Based on K-Means Clustering |
|
GUO Hui,HE Jie,CHEN Xiaohong
|
( Laboratory of Image Processing and Intelligent Information System, Wuzhou University, Wuzhou 543002, China)
|
| Abstract: In order to solve the problem of overly long time during fractal image coding,this paper focuses on a fast fractal coding algorithm based on K-means clustering.First of all,the variance method is employed to divide the range blocks into simple range blocks and complex range blocks;then,the K-means clustering algorithm is applied to classify the complex range blocks and domain blocks,and the nearest neighbor search approach is applied to search matching domain blocks,so as to match the corresponding range blocks with the domain blocks of the same type only within the neighboring scope.This method optimizes the searching process for matching blocks,thereby greatly shortening the encoding time.Test results show that,compared with the basic fractal coding algorithm,this method can increase the encoding speed by many times,with highquality reconstructed images. |
| Keywords: fractal image coding;K-means clustering;nearest neighbor search;variance method |
