子块均点特征分形快速图像压缩编码  

Sub-block Average Points Features Fast Fractal Image Compression Coding

作　　者：牛天婵 张爱华[1] 纪海峰[1] NIU Tianchan;ZHANG Aihua;JI Haifeng(School of Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China)

机构地区：[1]南京邮电大学理学院,南京210023

出　　处：《电讯技术》2019年第3期306-310,共5页Telecommunication Engineering

基　　金：国家自然科学基金资助项目(11471114; 61372125);江苏省自然科学基金项目(BK20160800)

摘　　要：分形图像压缩根据图像特有的自相似性,利用压缩仿射变换消除图像数据冗余度,进而实现图像压缩,实现较高的压缩比。然而,分形图像压缩编码具有计算复杂度高、运行时间过长的致命缺点,对于图像信息量巨大的当今社会来说不具有实用性。为解决基本分形压缩编码耗时过长的问题,提出了子块均点特征分形压缩编码算法,利用该算法将基本分形压缩编码的全搜索转为局部搜索,限定搜索范围,减少定义域块的搜索,在客观质量稍作牺牲的基础上加快了编码速度。将所提算法分别与五点和特征算法、1-范数特征算法、欧式比特征算法以及双交叉算法进行比较,仿真结果表明,在时间稍逊的情况下,所提算法在客观质量(Peak Signal-to-Noise Ratio,PSNR)上更优。Based on the self-similarity of images,the fractal image compression uses compression transformation to eliminate image data redundancy,thus realizing the image compression and achieving high compression ratio.However,for the fatal shortcomings of fractal image compression coding such as high computational complexity and long operation time,it is not practical for large image information of today′s society.To solve these problems,the sub-block average points features coding is presented to make the basic fractal compression coding from global search into local search,limit the search scope,reduce the search domain block,so as to speed up the encoding search.The simulation results show that the proposed algorithm is better than Five-Points Sum Algorithm,1-norm Algorithm,Euclidean Ratio Algorithm and Double Crossover Algorithm at the peak signal-to-noise ratio(PSNR)although the time is more.

关 键 词：图像压缩 子块均点特征 分形压缩编码 全搜索 局部搜索 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]