No Reference Image Quality Metric - An Integrated Approach

Image quality assessment becomes a standard practice in different image related applications because quality of the images play significant role to infer accurate decision. There are different approaches to assessing quality of images out of which most efficient one is developed based on human judgment, which don't require any reference image. Human judgment is usually expressed by linguistic variables, reflecting perception of human visual system regarding quality of an image.The aim of the dissertation work is to provide a computing platform to model human perception on assessing image quality without any reference image using fuzzy based methods. The work successfully deals with various issues and challenges towards development of an automated no reference image quality assessment method and designing quality metric comparable with state of the art quality metrics. In the course of the development process, several shortcomings of existing methods are addressed and solutions are proposed to achieve better and more promising performance.