提升小波变换的三种算法性能比较
|
摘要
离散小波变换(DWT)在语音,图像等信号处理中有着广泛的应用,在JPEG2000标准中就推荐采用5/3和9/7小波来分别进行无损和有损图像压缩,取代基于DCT变换的图像压缩,并且还推荐采用提升方法来实现。提出三种基于提升方法的二维离散小波变换的并行算法,并在超常超标量数字指令(VLIW)的数字信号处理器(DSP)上进行了性能方面的比较。这里,我们以在图像压缩中常用的不同分辨率的图像作为实验对象。实验结果表明,此三种算法对图像数据进行小波变换的处理时间有了明显缩短,并且实现在参数空间的不同点上都得到更好的效果。
Dimensional discrete wavelet transform(DWT) is becoming one of the standard tools for audio and image signal processor,the 5/3 and 9/7 filter-pairs included in JPEG-2000 was considered for lossless and lossy image compression,instead of based on DCT image compression and improved using lifting scheme.This paper proposes three arithmetic based on lifting scheme of two-dimensional discrete wavelet transform(2D DWT) and compare with respect to their performance on a very long instruction-word(VLIW) digital signal processor(DSP).Here,we choose various image sizes usually using in compression.The experimental results show that each realization can reduce the amount of time requirement and performs better for different points of the parameter space.
Dimensional discrete wavelet transform(DWT) is becoming one of the standard tools for audio and image signal processor,the 5/3 and 9/7 filter-pairs included in JPEG-2000 was considered for lossless and lossy image compression,instead of based on DCT image compression and improved using lifting scheme.This paper proposes three arithmetic based on lifting scheme of two-dimensional discrete wavelet transform(2D DWT) and compare with respect to their performance on a very long instruction-word(VLIW) digital signal processor(DSP).Here,we choose various image sizes usually using in compression.The experimental results show that each realization can reduce the amount of time requirement and performs better for different points of the parameter space.
引文
[1]Shapiro J M.Embedded image coding using zero trees ofwavelet coefficients[J].IEEE Trans.Signal Process.,1993(41):3445-3462.
[2]ISO/IEC FCD15444-1:2000.JPEG 2000 image codingsystem[EB/OL].March 2000.
[3]ISO/IEC JTC1/SC29/WG11,FCD 14496-1.Coding ofmoving pictures and audio[EB/OL].May 1998.
[4]Andreopoulos Y,Schelkens P,and Cornelis J.Analysis ofwavelet transform implementations for image and texturecoding applications in programmable platforms[J].Proc.IEEE Workshop on Signal Processing Systems,2001(1):273-284.
[5]Lafruit G,Nachtergaele L,Vanhoof B,and Catthoor F.Thelocal wavelet transform:a memory-efficient,high-speed architecture optimized to a region-oriented zero-tree coder[J].Integr.Comput.Aided Eng.,2000,7(2):89-103.
[6]Chrysafis C,and Ortega A.Line based,reduced memory,wavelet image compression[J].IEEE Trans.ImageProcess.,2000,9(3):378-389.
[7]Lafruit G,Nachtergaele L,Bormans J,Engels M,andBolsens I.Optimal memory organization for scalabletexture codecs in MPEG-4[J].IEEE Trans.CircuitsSyst.Video Techno.,1999,9(2):218-243.
[8]http://www.support.trimedia.philips.com
[9]SweldensW.The lifting scheme:a construction of secondgeneration wavelets[J].SIAM J Math Anal,1997,29(2):511-546.
[10]Taubman D.High performance scalable image compres-sion with EBCOT[J].IEEE Transactions on ImageProcessing,2000,9(7):1158-1170.
[11]Daubechies I,and Sweldens W.Factoring wavelettransforms into lifting steps[J].J.Four.Anal.Appl.,1998,4(3):247-269.
[12]阮秋琦,阮宇智.数字图像处理[M].北京:电子工业出版社,2003.
[13]李子萍.利用小波变换进行图像处理[J].临沧教育学院学报,2005,14(1):82-90.
[14]杨珂,刘明业.JPEG 2000标准下二维离散小波变换高速VLSI结构设计[J].北京理工大学学报,2005,1(5):394-398.
[15]邹江花,朱荣,秦前清.基于线扫式小波变换的海量地震数据准无损压缩算法[J].电子技术应用,2006(17):19-21.
[16]涂静波,李华刚,张云麟.Trimedia多媒体开发技术[J].重庆邮电学院学报,2002,4(13):51-55.
[17]李水根,吴纪桃.分形与小波[M].北京:科学出版社,2002:208-248.
[2]ISO/IEC FCD15444-1:2000.JPEG 2000 image codingsystem[EB/OL].March 2000.
[3]ISO/IEC JTC1/SC29/WG11,FCD 14496-1.Coding ofmoving pictures and audio[EB/OL].May 1998.
[4]Andreopoulos Y,Schelkens P,and Cornelis J.Analysis ofwavelet transform implementations for image and texturecoding applications in programmable platforms[J].Proc.IEEE Workshop on Signal Processing Systems,2001(1):273-284.
[5]Lafruit G,Nachtergaele L,Vanhoof B,and Catthoor F.Thelocal wavelet transform:a memory-efficient,high-speed architecture optimized to a region-oriented zero-tree coder[J].Integr.Comput.Aided Eng.,2000,7(2):89-103.
[6]Chrysafis C,and Ortega A.Line based,reduced memory,wavelet image compression[J].IEEE Trans.ImageProcess.,2000,9(3):378-389.
[7]Lafruit G,Nachtergaele L,Bormans J,Engels M,andBolsens I.Optimal memory organization for scalabletexture codecs in MPEG-4[J].IEEE Trans.CircuitsSyst.Video Techno.,1999,9(2):218-243.
[8]http://www.support.trimedia.philips.com
[9]SweldensW.The lifting scheme:a construction of secondgeneration wavelets[J].SIAM J Math Anal,1997,29(2):511-546.
[10]Taubman D.High performance scalable image compres-sion with EBCOT[J].IEEE Transactions on ImageProcessing,2000,9(7):1158-1170.
[11]Daubechies I,and Sweldens W.Factoring wavelettransforms into lifting steps[J].J.Four.Anal.Appl.,1998,4(3):247-269.
[12]阮秋琦,阮宇智.数字图像处理[M].北京:电子工业出版社,2003.
[13]李子萍.利用小波变换进行图像处理[J].临沧教育学院学报,2005,14(1):82-90.
[14]杨珂,刘明业.JPEG 2000标准下二维离散小波变换高速VLSI结构设计[J].北京理工大学学报,2005,1(5):394-398.
[15]邹江花,朱荣,秦前清.基于线扫式小波变换的海量地震数据准无损压缩算法[J].电子技术应用,2006(17):19-21.
[16]涂静波,李华刚,张云麟.Trimedia多媒体开发技术[J].重庆邮电学院学报,2002,4(13):51-55.
[17]李水根,吴纪桃.分形与小波[M].北京:科学出版社,2002:208-248.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心