Image denoising and image enhancing are two main technologies of image processing. In this thesis, image enhancing and denoising algorithms based on DT CWT(Dual-Tree Complex Wavelet Transform) are studied. A novel image enhancing algorithm based on transform domain processing and image domain processing is proposed first. Then a novel image denoising algorithm based on bivariate statistics is introduced. Since the processing of MRI image requires both the enhancement of the details and suppression of noise, a synthesized algorithm based on the above algorithms and edge detection is proposed.In proposed image enhancing algorithm, the high frequency wavelet coefficients are modified to enhance the local contrast and the details of image. Dual-tree Complex Wavelet Transform (DT CWT) is used in the wavelet transform due to its better directional selectivity. Then the overall contrast of the image is adjusted with nonlinear transforms in the image domain.In our image denoising algorithm, a parametric bivariate generalized Gaussian distribution (GGD) is used to describe the statistical distribution of DWT coefficients of the image. Then, based on maximum likelihood estimate(MLE), we can get the estimated parameters of the GGD. With these estimated parameters, Maximum A Posteriori (MAP) estimator can be used to restore the wavelet coefficients from the noisy observations.To achieve the detail enhancement and noise suppression at the same time, we apply the above algorithms to the same image to get an enhanced image and a denoised image. Then we combine those two images with the information got with edge detection. The processed image keeps good contrast at the edge while suppressing the noise in the smooth area. The result is quite satisfying visually.Finally, we developed a software which can be used to facilitate the study of image processing algorithms. The separation of the user interface and algorithm core makes the software maintainable and extensible while the algorithm core quite portable. |