Reseach On Vector Quantization Based On Wavelet Transform Image Coding System

Efficient compression of digital image information can not only promote the wideapplication of digital image,while areas of transmission over the Internet and Internetstorage media has a huge role in promoting the field. Therefore,for the multimediatechnology, network technology, computer hardware and software technology, digitalimage effective, quality compression has a very important significance.Vector quantization and wavelet decomposition technology, demonstrated a newfield of image compression applications. MATLAB software as an advanced softwaredevelopment platform, in particular its GUI tools for the software design and friendlyinterface provides a convenient platform, this article will vector quantization, waveletdecomposition techniques to MATLAB GUI tools of the three organic bonded togetherto design vector quantization of wavelet image codec system that is easy to operate,excellent performance, and friendly user interface.This paper describes in detail with rate-distortion adaptive vector quantization ofwavelet image codec design, given the principle of each module design and functionrealization, including: the whole image of the9/7wavelet transform, LLL thenormalized subband DPCM prediction, high pass subband quantization grid, latticecodebook design, the code word search, Huffman coding sub-band, high-pass subbandcoding method, decoding Finally this framework, in order to achieve the codec systemintegrity.Vector quantization of wavelet image coding system first reads the entire image, asshown in codec system visualization interface, according to the selected mode andwavelet transform of each sub-band quantization factor for image compression. After athree-layer wavelet transform10wavelet subbands using different quantization andencoding methods. For sub-band LLL DPCM predictive difference, then the differenceis Huffman coding, and for the high frequency sub-band with lattice vector quantizationmethod fixed length coding and Huffman coding method to compress the last encoderoutputs a binary stream file. By decoding module reads the encoded binary files, andfinally the peak signal to noise ratio (PSNR) and bit rate (bpp) calculations, adapteddecoding system also provides image rotation functions and the date, time functions.Experimental results show that the vector quantization described herein wavelet image coding system performance in image compression compared to conventionalimage compression algorithm performance has been greatly improved, and the system isdesigned in this paper has a friendly user interface, user-friendly operation.