Publication: Evaluations of quality measures for grayscale digital image
Date
2012-06-01
Authors
Lee, Chee Keng
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In this digital world of photography, it cannot be denied that almost everyone has a digital camera. It can either be a compact digital camera, digital SLR camera or even a mobile phone with camera. Therefore, it is very common occasion that images that are taken from a camera to be blurry or noisy. Many methods in image quality measures have already been developed and from all these methods, there are pros and cons of each measure. Hence an evaluation has to be done to compare these methods. Firstly, multiple images with different noise level, contrast level, blurring level and distortion are generated. Secondly, each methods of quality measure are implemented. There are as much as fifteen methods of evaluation that are implemented in this project. The next step is to do measurement on the images using the methods that are implemented. A human survey to determine image quality is done as well. Results of each measurement and human survey are collected and comparison is made. Finally, method with the best performance is determined. From the results obtained, certain methods can be proved to be suitable for certain degradation of images. With the comparison of the mathematical evaluations with the human survey, it is shown that the survey has an accurate result as well as the mathematical methods. As conclusion, universal quality index and structural similarity index show the best sensitivity in image quality measure.