摘要
在小波图像编码中,正交、双正交小波之所以经常被用于分解图像,是因为它们具有良好的去相关性能。从消失矩、正则性的角度分析这两种小波的特性,并加以对比。最后对不同频率特性的图像在嵌入式编码算法中使用这两类小波进行对比实验。结果表明:在相同的比特率时,使用双正交小波得到的重构图像质量较高,而且分解端具有较高的消失矩、重构端具有高正则性是双正交小波滤波器的设计标准。
In the field of image coding based on wavelet transforms, the orthogonal and biorthogonal wavelet are frequently employed to decompose image for their better decorrelation ability. This paper first gives a detailed investigation on them from the view of vanishing moment and regularity. Then according to frequency characteristics, all nature images are classified into 3 groups. Finally, in EZW algorithm, five classical wavelets are used tc decompose images, quantization and coding wavelet coefficients. Experiments results indicated that with the same bitrate the PSNR of reconstructed image obtained by biorthogonal wavelets is higher than that of orthogonal ones. When designing biorthogonal wavelet filter, high vanishing moment in decompositon filter and better regularity in reconstruction filter need to be well considered.
出处
《桂林电子科技大学学报》
2008年第3期203-206,共4页
Journal of Guilin University of Electronic Technology
基金
广西创新基金项目(2007105950701M04)
关键词
正交
双正交
消失矩
正则性
EZW
orthogonal
biorthogonal
vanishing moment
regularity
EZW
