A novel image compression based on the wavelet coding with the adaptive context quantization

摘要

In this paper, a novel image compression algorithm based on the wavelet transform is discussed. The wavelet transform operation is deployed firstly for the image waiting for being coded and the uniform quantization is used to reduce the spatial cost for the coding model subsequently. Then the coefficients of the wavelet transforming are decomposed into three parts. Three Context models are established for coding these partitioned coefficients separately. In purpose of finding the optimal Context quantization levels, the context quantization based on the minimum increment of the adaptive code length is proposed to tackle the "model dilution". The experiment results indicate that the proposed algorithm perform better compression efficiency than other predecessors. The proposed algorithm keep the acquisition of the optimal context quantization level adaptively with the minimum adaptive code length being achieved at the same time.