基于复指数矩的模糊聚类分形编码研究
|
摘要
在分形图像编码过程中,搜索每个R块的最优匹配D块所需要计算量相当可观,从而导致编码时间过长.图像像素块的复指数矩具有平移、旋转、缩放等多畸变不变性,这种多畸变不变性与图像的分形特性相契合,并且可以利用快速傅里叶变换实现.本文提出基于复指数矩和模糊聚类的快速分形编码方法,根据D块的复指数矩不变量利用模糊聚类对D块进行分类,进一步根据R块的复指数矩不变量寻找最优匹配的D块.实验表明,与其它方法相比,该快速分形编码方法在保持解码图像质量不变的同时,大大提高了分形编码的速度.
Traditional fractal coding has been widely applied to the image compression due to the high compression ratio.But the encoding in fractal image compressions are very time-consuming,because a large numbers of sequential search through a list of domains are needed to find the best match for a given range block.The Complex Exponent Moments(CEMs)are shift,rotation,scale and intensity distorted-invariant.This invariance can be used to match fractal image,and 2-D Fast Fourier Transform(FFT)algorithm is easily used to calculate CEMs.An effective fractal image compression based on CEMs and fuzzy clustering is proposed in this paper.Firstly,domain blocks are categorized using fuzzy c-mean-clustering approach.Then range blocks are compared to find the best domain blocks based on the CEMs.It shows in experimental results that the encoding is speed up with better performance in contrast with other fractal algorithms.
Traditional fractal coding has been widely applied to the image compression due to the high compression ratio.But the encoding in fractal image compressions are very time-consuming,because a large numbers of sequential search through a list of domains are needed to find the best match for a given range block.The Complex Exponent Moments(CEMs)are shift,rotation,scale and intensity distorted-invariant.This invariance can be used to match fractal image,and 2-D Fast Fourier Transform(FFT)algorithm is easily used to calculate CEMs.An effective fractal image compression based on CEMs and fuzzy clustering is proposed in this paper.Firstly,domain blocks are categorized using fuzzy c-mean-clustering approach.Then range blocks are compared to find the best domain blocks based on the CEMs.It shows in experimental results that the encoding is speed up with better performance in contrast with other fractal algorithms.
引文
[1] A.E.Jacquin.Image coding based on a fractal theory of iterated contractive image transformations[J].IEEE Trans.Image Process.1992,L(l):18-30.
[2]吴一全,郭建军,孙子翼.基于Krawtchouk矩不变量和核模糊聚类的快速分形编码[J].光电子.激光,2010,21(9):1412-1416.
[3]M.K.Hu.Visual Pattern Recognition by Moment Invariants[J]IRE Transactions on Information Theory1962,8(2):179-187.
[4] M.R.Teague.Image Analysis via the General Theory of Moments[J].J Opt Soc Am,1980,70(8):920-930.
[5] Y.Sheng and L.Shen.Orthogonal Fourier-Mellin moments for invariant pattern recognition[J].J Opt Soc Am A,1994,11(6):1748-1757.
[6] H.Ren,W.Wu,W.Bo,Y.Sheng,Z.Ping,Multidistortion-invariant image recognition with radial harmonic Fourier moments[J].J Opt Soc Am A Opt Image Sci Vis.,2003,20(4):631-637.
[7] Z.Ping,H Ren,J.Zou,Y.Sheng,W.Bo.Generic orthogonal moments:Jacobi-Fourier moments for invariant image description[J].Journal of Optoelectronics Laser,2007,40(4):1245-1254.
[8]姜永静.指数矩及其在模式识别中的应用[D].北京:北京邮电大学,2011.
[9]张莉,周伟达,焦李成.核聚类算法[J].计算机学报,2002,25(6):587-591.
[10]普运伟,金炜东,朱明等.核模糊C均值算法的聚类有效性研究[J].计算机科学,2007,34(2):207-210.
[2]吴一全,郭建军,孙子翼.基于Krawtchouk矩不变量和核模糊聚类的快速分形编码[J].光电子.激光,2010,21(9):1412-1416.
[3]M.K.Hu.Visual Pattern Recognition by Moment Invariants[J]IRE Transactions on Information Theory1962,8(2):179-187.
[4] M.R.Teague.Image Analysis via the General Theory of Moments[J].J Opt Soc Am,1980,70(8):920-930.
[5] Y.Sheng and L.Shen.Orthogonal Fourier-Mellin moments for invariant pattern recognition[J].J Opt Soc Am A,1994,11(6):1748-1757.
[6] H.Ren,W.Wu,W.Bo,Y.Sheng,Z.Ping,Multidistortion-invariant image recognition with radial harmonic Fourier moments[J].J Opt Soc Am A Opt Image Sci Vis.,2003,20(4):631-637.
[7] Z.Ping,H Ren,J.Zou,Y.Sheng,W.Bo.Generic orthogonal moments:Jacobi-Fourier moments for invariant image description[J].Journal of Optoelectronics Laser,2007,40(4):1245-1254.
[8]姜永静.指数矩及其在模式识别中的应用[D].北京:北京邮电大学,2011.
[9]张莉,周伟达,焦李成.核聚类算法[J].计算机学报,2002,25(6):587-591.
[10]普运伟,金炜东,朱明等.核模糊C均值算法的聚类有效性研究[J].计算机科学,2007,34(2):207-210.