Understanding the Effects of Visual Impairment on Visual Search

Dalal Aljasem, Andrew Howes

Abstract: The effect of vision loss on visual search has been extensively studied in psychology, but little is known about the relevance of these insights to human-computer interaction (HCI). This paper takes a first step towards a user model of visual impairments. Variants of the model are used to explain how people with Age-related Macular Degeneration (AMD), glaucoma or cataracts search for visual information on a computer display. The model formulates the visual search problem as a reinforcement learning (RL) problem. It is tested on its ability to predict the effects of visual impairments on how people search for objects in a crowded display or within a wide visual field. We discuss how the model might help the design of more accessible user interfaces and enhance user experience for those with vision impairments.