Assistive browser for conducting web transactions
Jalal Mahmud, Yevgen Borodin, I. V. Ramakrishnan · 2008
People with visual impairments use screen readers to browse the Web. Sequential processing of web pages by screen readers causes information overload, making web browsing time-consuming and strenuous. These problems are further exacerbated in web transactions (e.g.: online shopping), which involve multiple steps spanning several web pages. In this paper we present a lightweight approach for doing Web transactions using non-visual modalities. We describe how analysis of context surrounding the link coupled with a shallow knowledge-base with patterns and keywords can help identify various concepts (e.g.: "add to cart", "item description", etc.) that are important in web transactions. Our preliminary results show promise that presenting such concepts to the users can reduce information overload and improve their overall browsing experience.