Research Of The Vector Quantization Image Compression Coding Based On Wavelet Transform

With the information times coming,especially Internet has become more and morepopular. This makes people enjoy the convenience of the digital information society.Butbecause of the large amount of digital images,huge storage space and wide bandwidth arerequired. So saving storage space and reducing engrossing of transmission channel become ahotspot in research.The paper design a new image compression algorithm throughresearching the wavelet image coding and VQ theory, to further reduce storage space andreduce the transmission quantity.The paper further research the wavelet transform and VQ theory firstly, detailed introducedthe key technology of VQ,the weak point of the traditional codebook design method and the basic SOFMmethod. Secondly, detailed introduced all sorts of correlation of wavelet coefficients.Finally, Thisalgorithm is proposed: Firstly,according to the characteristics of wavelet coefficients,extract thevector using all kinds of relevance. combining the nonlinear interpolation method for vector quantization.In the code book design stage,use the improved SOFM network,selecting appropriate initial codebookcomprehensive use the constraint relation between Input vector and code words,the constraint relationbetween the code words,the absolute errors inequality delete algorithm.and the partial distortion deletealgorithm, reducing the number of codebook matching vector effectivly, improving the efficiency. Finallysimulates and realizes the algorithm on the matlab platform, contrast and analyses the performance ofthe new algorithm and the other correlation algorithms. The experiments show the methodenhancing the PSNR in the certain degree and improve the quality of compressed image.