Large scale flexible event-based clustering from photos in social media

  • Authors:

  • Mohamad Rabbath;Philipp Sandhaus;Susanne Boll

  • Affiliations:

  • OFFIS - Institute for Information Technology, Oldenburg, Germany;OFFIS - Institute for Information Technology, Oldenburg, Germany;University of Oldenburg Oldenburg, Germany

  • Venue:
  • Proceedings of the Third International Conference on Internet Multimedia Computing and Service
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

With the large amount of photos and other media items that are added daily to social media sites, it has become very hard for the user to identify the photos of a specific event, especially as it can be spread over a lot of users and albums. Additionally, people may have different definitions of what they see as an event, and in many cases this definition can change over time. In this paper we propose a method to enable users to easily browse photos in an event-oriented way. We exploit a probabilistic approach that performs the heavy computation offline, but flexibly leaves the definition of the event to the user in the online time. In our approach the user starts by an initial small set of photos which are related to an event or several events in her opinion, and retrieves the photos that are highly probable to belong to the same event. The user can interactively change this initial set based on the newly retrieved photos.