Theory Of Wavelet And Its Application In Video Compression

In the late 1980s, Daubechies proposed a systematical approach of constructing compactly supported orthogonal wavelet, and Mallat presented multiresolution analysis and discrete wavelet transform. At present, in the fields of wavelet, the lifting scheme is the key question, not only because of its generality and flexibility, but also because it allows efficient implementation. Signals and Image information are the important source for human being to explore the world outside and inside. How to use as few coefficients as possible to express the signals and image is the question needed to resolve in a lot of research fields. This paper mainly concluded the biorthogonal spline wavelet filter, studied on the theory of wavelet transform based on the lifting scheme, studied on some kinds of methods of motion estimation, proposed a new simple diamond search algorithm, studied the 3-Dwavelet coding algorithm, proposed a new algorithm—"3D SPECK" which fits 3-Dspace well and proposed a new concept "Virtual Decompose" using it in wavelet coding.We studied the theory of wavelet, and introduced the main work as follow:1, Because of its linear properties, the biorthogonal wavelet has wide application. Constructing wavelet filters using its character is the key question. In order to solve the coefficients of the fast lifting wavelet transform, a new method for calculating the coefficients of the lifting schemes was studied. Because the new method here does not use the Eucildean algorithm, it is easy to solve the coefficients. Moreover, its results are the same as that obtained by other methods. At the end of the chapter 2, we studied the statistic character of wavelet coefficient.2, Motion estimation is very complex. Diamond search algorithm has been succeed, but it has some fault. We proposed a new simple algorithm, which can judge from different dots.3, The 3D wavelet video coding is promising. We studied the 3D wavelet video compression, and proposed a 3D-SPECK coding method, which fits 3D wavelet video coding well.4, As the TV set becomes bigger and bigger ,the image of the video becomes bigger and bigger .When we decomposed them using wavelet, the LLband sometimes can as big as CIF. So, we proposed a new concept "Virtual Decompose" and used "virtual SPECK"in big image.