A Wavelet Transform Image Coding Algorithm Context Model To Quantify

With the development of computer software and hardware, effective image compression is becoming more and more important as the image is full of useful information. Statistical coding, transform coding, vector quantization coding, and predictive coding are some commonly used methods for image compression, among which, wavelet transform under transform coding has a higher compression ratio and better compression quality for image coding.This thesis is going to research the changes happened to wavelet-transformed images on the stream of Map and Sign by means of constructing Context model and quantitative model, and researching on compression. Traditionally, therefore solving "model cost" and achieving effective image compression is through organizing wavelet coefficients by reducing the model order or data range based on the theory of "Conditions of Reduced Entropy". This thesis puts forward a new method of effective image compression by means of increasing the order of the model and data range.When the conditional probability has been obtained by conducting hash algorithm to get effective statistics of the wavelet coefficients from of model, a mapping table can be finished through hierarchical clustering algorithm. The mapping table can effectively organize wavelet coefficients and solve the problem of "model cost" in Context model. Compared with the traditional method, two-pass encoding is more valid in image compression. This thesis also explores the versatility of the mapping table, which can be accepted in a certain field of application.