¿
DESCRIPTION (provided by applicant): Web browsing with assistive technologies such as screen readers and magnifiers can often be a frustrating and challenging experience for people with vision impairments, because it entails a lot of searching for content, forms, and links that are required for doing online tasks such as shopping, bill-payment, reservations, etc. This SBIR Phase II project will build on Phase I results and will continue the development and eventual deployment of Capti Screen Reading Assistant - a next-generation assistive technology, enabling goal- directed web browsing for people with visual impairments. With Capti Assistant, users will be able to stay focused on their high-level browsing goals that are expressed in natural language (spoken or typed). The Assistant will lead the users step-by-step towards the fulfillment of these goals by offering suggestions on what action to take at every step of the way and automatically executing the chosen action on behalf of the user. Suggested actions will include operations such as form filling, activating controls (e.g., clicking buttons and links), et. Capti Assistant will dramatically reduce the time spent by people with visual impairments on performing tasks online. The Assistant will significantly improve the speed and efficiency with which they can interact with the Web, thereby, making people with disabilities more productive in today's web-based economy. Given a user browsing goal, expressed in a natural-language form, Capti Assistant will utilize a predictive model to guide the user toward the goal. The unique
aspect of the Assistant is that its suggestions will be automatically learned from the user's own history of browsing actions and commands, as well as from the user's demonstration of how to accomplish browsing tasks that have not been done before. The Assistant will process user commands and present the suggested browsing actions to the user on demand, giving the user a choice between following the suggestions or continue browsing normally without accepting the suggestions. The functionality offered by the Assistant will go far beyond popular personal assistant applications such as Siri, which have not been designed specifically for people with vision impairments, and which cannot be used for ad hocweb browsing. Capti Assistant will go a long way towards bridging the growing web accessibility divide between the ways people with and without vision impairments browse the Web. For them, the Assistant will usher in a new era of independence and employability in our global web-based economy. Thus, from a broader perspective, goal- directed browsing will exemplify the vision of the Universally Accessible Web whose thesis is "equal access for all", i.e. anyone should be able to reap the benefits of the Web without being constrained by any disability.