Home | eMine: Web Page Transcoding Based on Eye Tracking Project Page
The World Wide Web (web) has moved from the Desktop and now is ubiquitous. It can be accessed by a small device while the user is mobile or it can be accessed in audio if the user cannot see the content, for instance visually disabled users who use screen readers. However, since web pages are mainly designed for visual interaction; it is almost impossible to access them in alternative forms. Our overarching goal is to improve the user experience in such constrained environments by using a novel application of eye tracking technology. In brief, by relating scanpaths to the underlying source code of web pages, we aim to transcode web pages such that they are easier to access in constrained environments. The project is supported by the Scientific and Technological Research Council of Turkey (TUBITAK) and is conducted in collaboration with the Web Ergonomics Lab (WEL) at University of Manchester. Project Objective Our objective is to use eye tracking data to generate an algorithm for identifying people's scanpaths and relating those scanpaths to elements of Web pages, such that Web pages can be transcoded to improve the user experience in constrained environments. In order to achieve that objective, eMine has four aims: Identify visual elements of Web pages automatically by using the underlying source code. In our previous work, we have developed a framework that can be used to identify visual elements of Web pages. This aim focuses on automating that process. Develop a novel algorithm to identify scanpaths in terms of visual elements of Webpages. "String-edit" algorithm is widely used to analyse people's scanpaths in given eye tracking data. Even though this algorithm has been successfully applied to Web pages, these studies use images of Web pages and none of them has investigated the relationship between scanpaths, visual elements of Web pages and the underlying source code. This aim focuses on developing a new algorithm based on "String-edit" that can be used to identify scanpaths in terms of visual elements of Web pages. Develop novel transcoding techniques based on our new algorithm. Even though various transcoding techniques have been researched, to our knowledge none of these techniques use eye tracking data to transcode Web pages. This aim focuses on developing novel transcoding techniques based on our new algorithm that identifies scanpaths in terms of visual elements of Web pages. Demonstrate that the proposed transcoding techniques improve the user experience in constrained environments. Even though we can technically transcode Web pages, we still need to demonstrate that the proposed techniques improve the user experience and therefore that is the focus of this aim. Visual Blocks Eye Tracking Fixations Transcoded Page
