The Research On Application Of Wavelet Network In Image Coding

Wavelet analysis is a rising mathematical analysis method, it has been applied successfully in many fields of engineering. This paper combines theory of wavelet analysis with conventional neural network. The neural network is formed by substituting wavelet function for activation function int classical neural network. It has ability of coherent approximation and L~2 approximation, and has very good ability of mapping and expansibility. These properties assure the wavelet neural network can be used in the filed of image coding.In this paper, wavelet neural network will be used in image coding. The priority of this method was discussed in theory. Recur to the theory of MRA, this paper constructs a wavelet neural network mode which can be used in image coding. The corresponding algorithm and formulas are presented. The results from simulated experiments show that this model can overcome the shortcomings of classical methods based on KL transform and bring out a better effect.Based on the research into the method named PCM, a multiple wavelet neural network model was constructed. The model has been used to improve the classical PCM in image coding and a better effect was brought out.This thesis consists of six chapters, and the main contents of each chapter are as follows.In chapter 1, we outline the background, meaning and the actuality of this thesis. Main results of this thesis will be introduced.In chapter 2, we outline the basic theory of image coding. Some common methods and international standards of image coding will also be introduced.In chapter 3, we introduced the development history and the basic theory of wavelet. The arithmetic of decomposing and reconstructing of wavelet named Mallat will be emphasized.In chapter 4, we introduced the development history and the basic theory of neural network. The design method of neural network's structure and the training arithmetic named BP Arithmetic will be emphasized. In chapter 5, we introduced the development history and the basic theory of wavelet neural network. On the one hand, we research into the design method of wavelet neural network's structure and the training arithmetic of wavelet neural network. On the other hand, we improve the image coding method based on KL transform by wavelet neural network, and the result is satisfying.In chapter 6, we first investigate the structure and training method of multiple wavelet neural network and then use it to improve the classical PCM method. The result is improving.