Live Demonstrator of EEG and Eye-Tracking Input for Disambiguation of Image Search Results

摘要

When searching images on the web, users are often confronted with irrelevant results due to ambiguous queries. Consider a search term like 'Bill': Results will probably consist of multiple images depicting Bill Clinton, Bill Cosby and money bills. Given that the user is only interested in pictures of money bills, most of the results are irrelevant. We built a demo application that exploits EEG and eye-tracking data for the disambiguation of one of two possible interpretations of an ambiguous search term. The demo exhibits the integration of sensor input into a modern web application.