NETWORKED EXPLORATION OF PERSONAL IMAGE COLLECTIONS

This paper describes our framework that enables a network of friends to effectively visualize and browse a shared image collection. The framework is very useful for geographically disconnected friends to share experiences. We present solutions for three tasks - (a) visualize what a particular user has done after (or prior) a specific event (b) visualize the event sequence that lead to two users to meet and finally (c) for a given user, determine interesting events. We solve the problem by defining an event model and deriving task-based media filter to determine the optimal media set. Then we develop two novel, complementary, visualization schemes - (a) spatio-temporal evolution and (b) event-cones. The first technique allows a person to follow the events for a specific group member, over space and time using maps as well as slideshows. The second technique shows a summary snapshot of all the events associated with the selected users allowing us to immediately understand their interactions over time. The user study shows statistically significant results thus indicating that the visualization environments work well.