MODEL-BASED IMAGE ENHANCEMENT FOR THE VISUALLY IMPAIRED

点击翻译按钮获取中文摘要

Enhancement of video images should enable the visual impaired greater access to TV. and other video sources. In previous years, contrast enhancement based on a band-pass filtering approach was developed and tested. Results were encouraging, showing improved performance in face recognition and in obtaining details from motion videos. Addition improvement will be gained by addressing other aspects of spatial vision beyond the loss of contrast sensitivity. We will study, in patients with central scotomas, the addition of wide-band feature markers improve the performance in localization, and in phase (shape) discrimination tasks, as suggested by studies carried out with normally sighted subjects. Using the results of these studies, we developed a novel extension of the previously used enhancement method, a wide-band enhancement that should have many advantages. The new method as a cartoon-like bar outlines (bipolar light and darkness) and increase the contrast of edge features to facilitate localization and discrimination. We will determine the optical parameters for such enhancement for patients with central scotomas, implement a software based approach to generate enhanced motion video segments and use such segments to measure the preference of visually impaired persons for the enhanced versions. We will continue to develop our visual model, and to use it in the development, evaluation and optimization of image enhancement for the visually impaired. The basic aspect of the study will examine changes in adaptation that occur in the peripheral visual processing with long term central visual loss. We will study performance differences on the number of visual tasks between patients with long term bilateral scotomas and with unilateral scotomas. The study of patients with monocular scotomas will provide better measure of the capability of the peripheral retina. The differences between monocular and bilateral scotomas patients performance will improve our understanding of adaptation to the force used of peripheral retina or form vision. Studies of long term facilitation low vision patients will be used in search for a measure of visual function that will explain and quantify deficit in peripheral retinal such as contour integration. The basic studies results will guide further fine tuning and optimization of the enhancement technique.