Journal of Systems Engineering and Electronics ›› 2011, Vol. 33 ›› Issue (8): 1875-1880.doi: 10.3969/j.issn.1001-506X.2011.08.37

• 软件、算法与仿真 • 上一篇    下一篇

基于最优分解和方向加权的静止图像压缩算法

杨春玲,曹端武,苏卓涵   

  1. 华南理工大学电子与信息学院, 广东 广州 510640
  • 出版日期:2011-08-15 发布日期:2010-01-03

Still image compression algorithm based on optimal decomposition and directional weighting

YANG Chun-ling, CAO Duan-wu, SU Zhuo-han   

  1. School of Electronic and Information Engineering, South China University of Technology, Guangzhou 510640, China
  • Online:2011-08-15 Published:2010-01-03

摘要:

基于小波变换的JPEG2000是新一代静止图像压缩标准,具有很多优点。但是由于二维小波是一维小波的张量积,它只有有限的方向,因而不能较好地捕捉图像的方向信息。基于方向滤波的小波contourlet变换(waveletbased contourlet transform, WBCT)可以弥补这个不足,它比小波变换能更稀疏地表示图像信息。把WBCT嵌入在JPEG2000框架中,并结合其性质,提出了基于最优分解和方向加权的静止图像压缩算法。实验结果表明,该算法在中低码率下可以更好地表现图像的轮廓和强纹理信息,在高码率下也能较好地编码图像的弱纹理信息。不管是参照客观质量评价方法,还是凭主观视觉感受,该算法的压缩性能均优于JPEG2000。

Abstract:

JPEG2000 based on wavelet transform is the next generation still image coding standard. It has many advantages. But the two-dimensional wavelet is just the tensor product of one-dimensional wavelet. It can not capture the image direction information well because of its lack of directions. Waveletbased contourlet transform (WBCT) based on directional filter can make up this deficiency. Compared with the wavelet transform, WBCT can more sparsely express image information and is embedded in the JPEG2000 framework. And a new still image compression algorithm, based on optimal decomposition and directional weighting is proposed. The experimental results show that the algorithm has better effect on contours and strong texture information of images in low bit-rate and weak texture information in high bit-rate, and it has better compression performance than JPEG2000 whether on objective evaluation methods or subjective visual feeling.