Integer Dct Transform Research And Application In Image Lossless Compression

The image compress includes loss compress and lossless compress. The dissertation mainly studies image lossless compress and its application. Along with progress of society and perfect of legal system, the image lossless compress has been currently widely used in the field of the authentication underwriting image, the archives image and the medical image, so it has practical significative for us to do an in-depth study on image lossless compress. The dissertation combine the theory, with using integers DCT transform to realize grey image lossless compress with the method of reversible color space integers transform and reversible integers DPCM prediction to realize lossless compress from grey image to color image with Huffman Coding Method via Visual C++ program. DCT is based on FFT (Fast Fourier Transform), so integers FFT transform must be structured firstly. The wavelet transform has been altered to the wavelet transform that maps integer to integer with lifting theory by Daubechies. We apply this theory to FFT and DCT. Lifting matrixes is the base of integers transforms, so structuring lifting matrixes is the emphasis in FFT and DCT integers transform. First, employing ordinary butterfly operations formula in FFT, lifting matrixes has been got and FFT transform that maps plural integer to plural integer has been achieved. Then, according to characteristic of four kinds of butterfly operation in real number signal integer FFT transform keeping modulus conjugating with symmetry, different lifting matrixes has been got and real number signal integer FFT transform keeping modulus conjugating with symmetry can been achieved. In the end, according to the relation of real number signal FFT and DCT, integer DCT transform can been got. In this paper, because direct current coefficients of 8×8 block have been predicted with integer DPCM , each steps of transform in DCT and DPCM were made up of the reversible integer transform. So, image can be renewed losslessly. Integer DCT transform is also used in the gradual image compress. Cuttingthe amount of 8×8 block high frequency coefficient of integer DCT transform to realize the gradual image compress from vague to clear-cut.In this paper, the entropy and the compress ratio of integers DCT transform were compared with ones of traditional DCT transform and integer DPCM prediction, and data characteristic of integer DCT transform was analyzed. Test result proved that it is reasonable to use integers DCT transform in the image lossless compress.The result of experimentation that the archives images and historical materials images were compressed with the program of this paper shows that it has valuable to use integers DCT transform in the archives images and historical materials images lossless compress.