四线和特征的快速分形图像编码
|
摘要
分形图像编码具有压缩比高、解码速度快、重构图像质量高等特点,但因这种算法在编码时定义域的搜索量十分巨大,导致其计算复杂度高、编码时间过长,阻碍了它的实用性和普遍应用。为解决此问题,文中提出一种基于四线和特征值编码算法,该算法根据匹配均方根误差与四线和特征间的关系,将全局搜索转化为局部搜索(近邻搜索),限定搜索空间,减少定义域块的搜索,从而提高编码速度。仿真实验结果表明:该算法解码图像质量在客观上优于1-范数特征算法;与基本分形编码算法相比,基于四线和特征算法在主观上不改变重构图像质量,但在编码速度上却得到极大提高。
Fractal image encoding has many advantages,such as high compression ratio,high decoding speed and high quality of reconstructed image. However,its high computational complexity and long encoding time make it impractical because of the huge amount of search in the domain. Aiming at these problems,an encoding algorithm based on the sum of four lines eigenvalues is proposed. And,the relationship between the matching error and the four lines eigenvalues converts global search to local search( neighbor search) to limit the search space and speed up the encoding. The simulation experiment results show that the objective decoding image quality by the algorithm of this paper is superior to 1-norm features algorithm. Compared with the basic fractal coding algorithm,it doesn't change the subjective quality of reconstructed image but the speed of encoding has been greatly improved.
Fractal image encoding has many advantages,such as high compression ratio,high decoding speed and high quality of reconstructed image. However,its high computational complexity and long encoding time make it impractical because of the huge amount of search in the domain. Aiming at these problems,an encoding algorithm based on the sum of four lines eigenvalues is proposed. And,the relationship between the matching error and the four lines eigenvalues converts global search to local search( neighbor search) to limit the search space and speed up the encoding. The simulation experiment results show that the objective decoding image quality by the algorithm of this paper is superior to 1-norm features algorithm. Compared with the basic fractal coding algorithm,it doesn't change the subjective quality of reconstructed image but the speed of encoding has been greatly improved.
引文
[1]RIBARIC S,FRATIC I.A biometric identification system based on eigenpalm and eigenfinger featyres[J].IEEETrans on PAMI,2005,27(11):1698-1709.
[2]李强,裘正定,孙冬梅,等.指横纹:一种新的生物身份特征[J].自动化学报,2007,33(6):596-601.
[3]何传江,蒋海军,黄席樾.快速分形图像编码的一种特征方法[J].电子学报,2004,32(11):1864-1867.
[4]何传江,申小娜.改进分形图像编码的叉迹算法[J].计算机学报,2007,30(12):2156-2163.
[5]LAI C M,LAM K M,SIN W C.Improved searching scheme for fractal image coding[J].Electronics Letters,2002,38(25):1653-1654.
[6]何传江,杨静.基于形态特征的快速分形图像编码[J].中国图像图形学报,2005,10(4):410-414.
[7]HARTENSTEIN H,SAUPE D.Lossless acceleration of fractal image encoding via the fast Fourier transform[J].Signal Processing Image Communication,2010,16(4):383-394.
[8]袁宗文,鲁业频,杨汉生.半叉迹特征的快速分形图像编码[J].计算机工程与应用,2016,52(3):197-201.
[9]LEE C K,LEE W K.Fast fractal image block coding based on local variances[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,1998,7(6):888-891.
[10]李高平,何传江,黄娟娟.提高分形图像编码质量与速度的方案[J].计算机仿真,2006,23(5):163-166.
[11]何传江,李高平.分形图像编码的改进算法[J].计算机仿真,2004,21(8):62-65.
[12]李高平.主对角和特征的快速分形图像编码[J].计算机工程与应用,2010,46(26):176-178.?
[2]李强,裘正定,孙冬梅,等.指横纹:一种新的生物身份特征[J].自动化学报,2007,33(6):596-601.
[3]何传江,蒋海军,黄席樾.快速分形图像编码的一种特征方法[J].电子学报,2004,32(11):1864-1867.
[4]何传江,申小娜.改进分形图像编码的叉迹算法[J].计算机学报,2007,30(12):2156-2163.
[5]LAI C M,LAM K M,SIN W C.Improved searching scheme for fractal image coding[J].Electronics Letters,2002,38(25):1653-1654.
[6]何传江,杨静.基于形态特征的快速分形图像编码[J].中国图像图形学报,2005,10(4):410-414.
[7]HARTENSTEIN H,SAUPE D.Lossless acceleration of fractal image encoding via the fast Fourier transform[J].Signal Processing Image Communication,2010,16(4):383-394.
[8]袁宗文,鲁业频,杨汉生.半叉迹特征的快速分形图像编码[J].计算机工程与应用,2016,52(3):197-201.
[9]LEE C K,LEE W K.Fast fractal image block coding based on local variances[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,1998,7(6):888-891.
[10]李高平,何传江,黄娟娟.提高分形图像编码质量与速度的方案[J].计算机仿真,2006,23(5):163-166.
[11]何传江,李高平.分形图像编码的改进算法[J].计算机仿真,2004,21(8):62-65.
[12]李高平.主对角和特征的快速分形图像编码[J].计算机工程与应用,2010,46(26):176-178.?