Research On Wavelet Based Contourlet Transform And Its Application On Image Processing

Wavelet based contourlet transform is a new kind of image representations whichis of multi-resolution, local supporting and multi-direction. The first stage of thetransform provides subband decomposition using wavelet transform instead oflaplacian pyramid in the contourlet transform to remove the redundancy of thelaplacian pyramid. And the second stage is a directional filter bank which producesangular decomposition at the same directions. Wavelet based contourlet transformbrings a more sparse image expansion and can capture the intrinsic geometricalstructure of images, which is more widely applied in image processing than wavelettransform and contourlet transform.The principle and implementation of wavelet based contourlet transform and itsapplication on image fusion and image enhancement is systemically studied in thispaper. The main work and achievements are as follows:1. The theory, implementation of 2-D separable wavelet transform wasparticularly studied. The disadvantage of 2-D separable wavelet transform in imageprocessing and multi-scale geometrical analysis was particularly discussed. Thetheory, implementation of contourlet transform and wavelet based contourlettransform was particularly studied. The precedence of contourlet transform towavelet transform and the precedence of wavelet based contourlet transform tocontourlet transform was illustrated through non-linear approximation experiments.2. The theory of multi-focus image fusion was particularly studied including itsconcepts and implementation algorithms. The design and the selection of the fusionrule was particularly researched and discussed. A method for multi-focus imagefusion based on wavelet based contourlet transform was proposed taking advantageof the transform's good property of multi-resolution and multi-direction. Combingthe local energy matching rule, the method got better effect than wavelet transform and contourlet transform, and decreased the RMSE value.3. The theory of image enhancement was particularly studied including itsconcepts and implementation algorithms. The design and the selection of theenhancing functions was particularly researched and discussed. A method for imageenhancement based on wavelet based contourlet transform was proposed takingadvantage of the transform's good property of multi-resolution and multi-direction.Combing A. F. Laine's enhancing function, the method got better effect than wavelettransform and contourlet transform to enhance the details of images.4. Shift-variance in wavelet based contourlet transform caused by criticaldown-sampling in the first stage wavelet transform results in visual artifacts called"Gibbs-like" phenomena. Cycle spinning was introduced to decrease theshift-variance and restrain the distortion and improved the image' quality.