Detail and contrast enhancement in images using dithering and fusion

Abstract

This thesis focuses on two applications of wavelet transforms to achieve image enhancement. One of the applications is image fusion and the other one is image dithering. Firstly, to improve the quality of a fused image, an image fusion technique based on transform domain has been proposed as a part of this research. The proposed fusion technique has also been extended to reduce temporal redundancy associated with the processing. Experimental results show better performance of the proposed methods over other methods. In addition, achievements have been made in terms of enhancing image contrast, capturing more image details and efficiency in processing time when compared to existing methods. Secondly, of all the present image dithering methods, error diffusion-based dithering is the most widely used and explored. Error diffusion, despite its great success, has been lacking in image enhancement aspects because of the softening effects caused by this method. To compensate for the softening effects, wavelet-based dithering was introduced. Although wavelet-based dithering worked well in removing the softening effects, as the method is based on discrete wavelet transform, it lacked in aspects like poor directionality and shift invariance, which are responsible for making the resultant images look sharp and crisp. Hence, a new method named complex wavelet-based dithering has been introduced as part of this research to compensate for the softening effects. Image processed by the proposed method emphasises more on details and exhibits better contrast characteristics in comparison to the existing methods.

Date of Award	2018
Original language	English