الفهرس | Only 14 pages are availabe for public view |
Abstract ABSTRACT This thesis deals with the issue of quality enhancement of digital images. Both contrast and resolution enhancements are considered simultaneously. Contrast enhancement is performed with histogram equalization (HE) and Contrast Limit Adaptive Histogram Equalization (CLAHE). The CLAHE proved to have the best performance due to its adaptive nature. The resolution enhancement is performed with interpolation techniques. Both polynomial and inverse interpolation techniques are considered and compared. The main advantage of the inverse interpolation is that the low resolution degradation model is inverted in the interpolation process. Four hybrid enhancement models are presented in this thesis based on the selection of the CLAHE and interpolation schemes. Simulation results reveal that the inverse interpolation in conjunction with the regularized interpolation achieves the best performance. This is attributed to the constraints on the inverse solution that preserve smoothness of flat areas and sharpness of edges and also to the adaptive nature of the CLAHE.There are no reference images to measure the quality with in the assessment process. Therefore, there is a need to use no-reference quality metrics in the assessment. We will adopt some metrics for this purpose including average intensity, contrast, spatial, and spectral entropies |