Abstract
Objective image quality measures play an important task in various image processing applications. The object of this dissertation is to develop new methods for medical image quality measurement. The methods are based on a statistics method, the Moran I statistics. Since a higher Moran Z value means that more structured information exists in the image and random noise is less likely in the image. So, the Moran Z value could be used to represent the spatial properties of mapped data. Image manipulation, such as compression, filtering and denoising, results in a variation of spatial correlation. Using of Moran I statistics to evaluate the variation of spatial information is a good choice. The peak ratio of Moran Z histogram between the manipulated and original images is used as an index for degradation or blurring after image compression or filtering. It can also be as an image blurring and compression index when applied to evaluate images in various compression ratios or filtering methods. A novel quality index is developed by applying Moran I statistics to various processed images and the measured values correlate well with the degree of quality degradation. In addition, the Moran Z value with a statistical test proved that the difference between original and manipulated images for subtle differences in them is attributed to chance. The Moran I statistics is a powerful for measuring image quality and this method can apply to many aspects of image quality evaluation.