Combining. Rough Sets And Neural Networks For Image Vector Quantization Coding Research

This paper analysis the status and the development trend of the image compression and summarize the progress status of the image compression around the world. And introduced some theoretical basic, rules, relevant concepts and three important technique of vector quantization. Have a research on image vector quantization compression method for image, which is the vector quaantization using self-orgaanization-feature-map. And the paper proposes an improved SOFM vector quantization method. All these are implemented using the MATLAB simulation tools. A deep research is focused on the classify vector quantization based on the rough sets and SOFM. The feature to classify the image block is following: frequency feature in DCT-domain and the statistics feature in spatial-domain. All the three classified VQ method is implemented and simulated using MATLAB. And the experimental results is given and is discussed.