Automatic Profile Generation for Visual-Impaired Users

Abstract

Currently, there are a growing number of tools that allow web developers to evaluate the accessibility of their web pages and sites. Many tools also prompt the developer to make specific repairs; and some tools automatically follow links to evaluate multiple pages within a site or within an entire domain. Although tools such as these can be very useful to identify accessibility problems, many accessibility issues are subjective and cannot be assessed without manual inspection. We believe that accessibility assessment and repair should be addressed as related activities intelligently supported. To do so, a main aspect to be considered is the automatic identification of user disabilities. In this paper we introduce an agent-based solution to tackle this problem. The agent is conceived to provide knowledge for the creation and classification of visually-impaired users profiles in terms of stereotypes. We apply and test the performance of our agent by profiling surveyed users. Our solution will be part of a multi-agent system to drive intelligently the accessibility conformance process.