说明：  小波变换源码，国外VC开发的，效果比较好 I ve been involved with wavelet-analysis since my Ph.D studies and over the years developed various wavelet-transforms C++ libraries. This one concerns 2D implementation of the Fast wavelet transform (FWT). The 2D FWT is used in image processing tasks like image compression, denoising and fast scaling. I intended to design the implementation of the 2D FWT with custom filter support and simple in usage. I ve equipped it with known filter families: Daubechies, Coiflets and Biorthogonal wavelets. The wavelets used in compression tasks are Bior53 and Bior97. With Daub1 filter you can achieve fast dyadic image scaling algorithm. You may also experiment with other filters for compression and denoising and compare the results, you may also add your custom wavelets. The nice feature of this library is that reconstruction FWT filters are rearranged the way that you perform just convolution operations on the spectrum, which is very important if you d like MMX
(err)