Model-Based Image Enhancement for the Visually Impaired

Schepens Eye Research Institute, Boston MA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Enhancement of video images should enable visually impaired people greater access to TV and other video sources. In previous years, contrast enhancement based on band-pass filtering and wide-band enhancement approaches were developed and tested. Results were encouraging, showing improved performance in face recognition and in obtaining details from motion videos as well as modest preference for enhanced images in both static and motion video. We believe that the modest level of preference for enhancement was a result of short-term adaptation to enhanced images, an effect that was recently reported from a number of labs. We will investigate these adaptation effects, including those in patients with central field loss, and will test enhancement methods that aim to overcome these effects. Also, we will continue evaluation of improved versions of the wide-band enhancement and the MPEG-based enhancement, both of which could benefit also from the same counter-adaptive approaches. The basic vision modeling aspect of the study will examine changes and adaptations that occur in peripheral vision processing with long-term central visual loss. In particular, studies of long-range facilitation in visually-impaired patients will be used in a search for measures of visual function that will explain and quantify deficits in pattern vision in peripheral retina. Studies of temporal aspects of vision started in the last grant period will be continued and expanded. Results of basic studies will guide further refinement of the visual model, and tuning and optimization of the enhancement techniques. Through image enhancement of video-based images, we expect to improve the quality of life of people with visual impairments.