The Research And Development Of Zero-tree Image Coding Algorithm Based On Wavelet Transform

The purpose of our research is to study the data compression coding method of still image. In this paper, an improved Zero-tree coding method based on wavelet transform is proposed. With the related experience, it was proved effective.In recent years, the research of Image coding method has become one of the most active fields in information technology by way of the imp of communication, medium store, multimedia computer technology, etc. Especially in 21st century, with the development of electronic and communicate technology, it is possible to realize the Video telephone, meeting TV, signal TV, information high way and etc. On this occasion, it inevitably becomes one of the main tasks to seek after the effective image coding method.This paper can be divided into six parts. Firstly, the background and significance of our research is introduced. Secondly, some basic knowledge of image coding technology are discussed. Thirdly, the theories of wavelet transform in image compression were described. With the analysis on distribution of the transform coefficient, it is proven that zero-tree structure is the best choice. Eventually, the advantage and shortage of the traditional Zero-tree coding method is pointed out. Based on it, a new algorithm of image coding is developed and some testing results are given to testify the efficiency of the presented algorithm.The advanced algorithm consists of three parts. By choosing more suitable Biorthogonal wavelet basis and symmetric periodic extension to improve the compression quality. Since most energy focus in the low frequency part, a little distortion will effect the quality of entire image heavily. So this part is dealt with separately. DCPM method is employed to improve the quality of image coding with making full of the relativity of the wavelet coefficients. According to the different importance degree and vision character of the high frequency wavelet coefficients, different thresholds are applied toremove small coefficients; Then Mannos module are applied to remove the visual redundancy; In order to produce more zero-tree, horizontal and vertical wavelet coefficients are transformed accordingly to improve the compression ratio.