Coding Techniques Based On Balanced Multiwavelets

Technologies for digital image compression have been the subject of much study over the past decade. Methods based on wavelet transform developed rapidly because of its better properties, bi9/7 and bi5/3 wavelet was adopted in existing image compression standards—JPEG2000. Recently the theory of multiwavelets was considered a generalization of wavelet, it can combine all desirable transform features, such as symmetry, support and orthogonality, and has more good characteristics. This made many researchers study on it. But its filter is matrix filter, the scalar input signal must be pre-filtered to vector signal. Generally pre-filter was designed bad, 'balanced' voided pre-filter, how did the balanced multi-filter features influenced image compression ? Because of the special sub-band structure of multiwavelets transform, how to combine SPIHT to it was another critical questions.According to these problems mentioned above, this thesis gives some research on the construction of balanced multiwavelet systems and the application of balanced multiwavelets to image compression. The main work: designed sa4bal balanced multiwavelets, A comprehensive analysis of balanced multi-filter and its effect on image compression; The analysis of the statistic features of image based on balanced multiwavelets transform, provided reference to the application of it; furthermore, two quantization methods modified according to SPIHT based on scalar wavelet, the algorithm made use of the correlation of coefficients of balanced multiwavelets, rearranged zero tree structure, and enlarged the depth of orientation tree and the possibility of zero tree's appearance, it improved PSNR, and analysis of their applicability on different balanced multiwavelet; many different image was selected to test our anlysis, the result of simulation shows our methods have better performance, at the same time, it proved the application of balanced multiwavelets to image compression has great potential.