Output details
11 - Computer Science and Informatics
Queen Mary University of London
A semi-automated approach to balancing of bottom-up salience for predicting change detection performance
<23>In this paper we present the first ever artificial intelligence driven technique for generating experimental change blindness test image pair; previously all experimental sets had been created manually. This work was featured by the BBC (http://www.bbc.co.uk/news/10284925) and numerous media outlets worldwide; the work was presented at the Secret Cinema science festival 2010 (600 people took the test) and Royal Society Summer Exhibition 2009 (http://royalsociety.org/summer-science/2009/magic-of-computer-science/) 5000 visitors. It is now the basis for a popular smartphone app with over 1500 downloads. We are working with marketing Agency Incahoots in applying the technique to aid package design debbie@debbiesimmons.com.