Still Image Encoding And Decoding Algorithm Based On Wavelet Transform And Realization

Wavelet analysis is a emerging studies developing quickly during recently ten more years. It has deep theories and been used broadly. It is the epoch-making result of developing of Fourier analysis.Wavelet discussed in this paper is a basis of L2 or other spaces created by a function (mother wavelet) expand and contract and shift. General functions may be represented as Wavelet series. A kind of stabile, efficient and fast algorithm of Wavelet transform may be gotten through multi-resolution analysis. At the same time because of localization of time domain (spatial domain) and frequency domain Wavelet transform is superior to Fourier transform.Compression and decompression of still images using wavelet transform is discussed in this paper. Images firstly are transformed from spatial domain to frequency domain through wavelet transform. Because of the multi-resolution analysis property, after wavelet transform original images will be transformed into a number of sub-images in different resolution ratio. Among them sub-image in the low resolution ratio is approximation of original image and it include most energy of original image. While sub-images in the high resolution ratio are details of original image. This progress is called image wavelet decomposition. Reverse wavelet transform put these sub-images together so that original image is recovered. This progress is called image wavelet reconstruct.Distribution of wavelet coefficients in those sub-images is just appropriate for the properties of human vision systerm. It is that eyes are most sensitive to sub-image in the low resolution ratio and less sensitive to sub-images in the high resolution ratio. So a little loss of information in high sub-images may not bring out marked change to vision effect of construct image. Then it is possible to realize high compression ratio of images. In this paper the principle of which wavelet transform can be used in images compression is discussed on basis of statistics and analysis of image's wavelet coefficients after wavelet decomposition. Also some kinds of quantify and coding schemes are discussed including Scalar Quantization, Vector Quantization, Embedded Zerotree Wavelets Encodings Run Length Coding, Huffman Coding and so on. Among them Scalar Quantization and Embedded Zerotree Wavelets Encoding are the important point.At last of the paper the principle of wavelet chip using for image compression and decompression produced by American ADI company is discussed. A mature scheme of image compression and decompression is introduced and a software is made to simulate the scheme.