摘要
研究采用了提升构造参数化的双正交小波类方法,构造了一类仅用一个自由参数表达的双正交插值小波,并推导出了其相关小波滤波器的参数表达式。自由参数为小波的优化设计提供了一个自由度,以编码增益为优化准则,构造出一种新的插值小波,其滤波器系数为二进制分数,可实现非乘法运算的离散小波变换(DWT)。实验结果表明,其图像压缩性能超过在小波变换编码领域中应用最为广泛的9/7小波(CDF-9/7),而且计算复杂度降低了17%以上,因此,在压缩性能和计算复杂度两者之间取得了一个较好的平衡。
This paper depicts how to construct parametric biorthogonal wavelet family via lifting scheme, and gives the complete construction of a new class of parametric biorthogonal interpolating wavelets with one parameter. The exact parameter expressions of their associated interpolating filter banks are also derived. The free parameter provides a degree of freedom to optimize the resulting wavelets, and a previously unpublished interpolating wavelet is obtained with respect to the coding gain criteria, which has binary filter coefficients and can realize a multiplication-free discrete wavelet transform （DWT）. Simulations show that the new wavelet has exhibited image compression performance superior to the most widely used 9/7 tap wavelet by Cohen et al. in the field of wavelet transform coding, yet its computational complexity has decreased by more than 17%. This indicates a better tradeoff between compression performance and computational complexity.
出处
《中国图象图形学报》
CSCD
北大核心
2010年第4期557-564,共8页
Journal of Image and Graphics
基金
国家自然科学基金项目(60673087)
江苏省高校自然科学基础研究项目(07KJD520005)
关键词
双正交小波
插值滤波器
离散小波变换
提升
压缩性能
计算复杂度
biorthogonal wavelet, interpolating filter, discrete wavelet transforms （DWT）, lifting scheme, compression performance, computational complexity
