STUDIES OF LOW VISION READING AND FACE RECOGNITION

患有中央视野丧失的低视力受试者阅读速度要慢得多

Low-vision subjects with central field loss read much more slowly than would be expected on the basis of standard clinical measures of visual function (e.g. visual acuity and contrast sensitivity). The principal goal of the proposed research program is to determine what limits reading performance in this group of subjects and to explore new techniques of Image enhancement to circumvent the visual limitations on reading in the absence of central vision. We will study reading performance in young and elderly normal observers and in low-vision subjects with central scotomas due to macular disease, primarily age-related maculopathy. To determine the limitations on reading of normal peripheral vision, normal observers will read text in the presence of an artificial scotoma stabilized on the retina with a dual-Purkinje eye tracker. Since inappropriate eye movement may severely limit peripheral reading, we will also study reading performance by presenting the text one word at a time in a fixed retinal location. This technique, known as rapid serial visual presentation, or RSVP, virtually eliminates the need for saccadic eye movements. We will determine the extent to which the peripheral retina of low vision subjects behaves like normal peripheral retina. We will measure spatial phase or position sensitivity using phase discrimination and vernier tasks. Lateral making ("crowding") effects will be investigated by measuring letter recognition in the presence or absence of flanking patterns, and metacontrast masking (backwards and forwards masking in the temporal domain) will be measured with a word recognition paradigm. Spatial undersampling will be investigated by comparing detection and orientation discrimination of grating patches, and temporal undersampling will be evaluated by comparing CFF with two- pulse discrimination. In each case we will the compare performance of low-vision subjects with central scotomas to that of normal subjects using eye-tracker controlled peripheral viewing. We will determine the relation of low-vision reading ability to the size, location, and density of central scotomas, and any discovered abnormalities of peripheral visual function. Given a better understanding of the visual limitations on peripheral reading performance, we will investigate new techniques for enhancing or modifying the image in order to improve performance. We will explore non-linear magnification and increased letter spacing to overcome crowding problems, low-pass spatial and temporal filtering to reduce lateral and metacontrast masking, and the use of static and dynamic texture to improve letter segregation.