Gaze Based Learning and Access for Search Engine
Dinesh Madhavrao Bodewar, Yogesh M. Bande, Prasad Laxmikant Telkikar, Sayali S. Meshram · SSRN Electronic Journal · 2015
Human computer interaction is often frustrating for the user. Donald Norman described gulf of evaluation as the difference between user’s expectations and the system’s response. When the search returns hundreds of results to a query, the user is required to spend a lot of time in finding the pages of his interest. A solution to this problem is not easy. Researchers have attempted to capture the context of user’s attention as he scans through the display. Techniques based on implicit feedback are being employed for this purpose. Implicit methods involve observing user’s action and environment, analyzing the collected data and inferring what might be relevant to him. In our proposed system we are trying to capture user’s interest through implicit feedback technique. We are using an ordinary web camera and a customized web browser in the setup. This technique monitors user activities as he browses through the web. It keeps track of the time user spends on a webpage. Most importantly we are attempting to collect the information related to gaze and eye movements of the user. Our intention is to detect with accuracy the topic under the gaze. We will then correlate the collected information with the content of the web page and assign some weight to it. Periodically the system will analyze the collected information and build a list of highly probable items of interest. We will then try to generate inference rules for each user. Inference rule can be shown to the user to get his confirmation. Finally the system will attempt to determine the interesting pages from the searched results whenever the user searches the web. In this way we intend to enhance user experience and reduce the gulf of evaluation.