Development of image display-specific evaluation and optimization schemes.

摘要

An electronic display is typically the last stage of an imaging chain before the human observer. Some of its physical properties can make it a bottleneck in the chain. This dissertation is an attempt to study and cope with certain critical properties of electronic displays, specifically spatial resolution, noise and contrast. The focus of this study lies in two parts. One is the measurement and estimation of display properties, including the spatial resolution and the spatial noise. The other is on the development of display-based image processing algorithms, which can compensate for the degradations from the specific display. The basic idea is to perform operations on the digital image before it is displayed, which will counteract the degradations that come from the action of the display systems. The properties of the display will be derived mainly from the digital images taken by a high quality CCD camera on the display. Based on the display properties estimated, several image processing techniques have been developed in this project to optimize the spatial resolution, noise and contrast resolution of the specific display under consideration. We have also developed a novel error-diffusion-based algorithm for adaptive image contrast enhancement which incorporates the properties of the display as well as the properties of the image.