Annotation suggestion and search for personal multimedia objects on the web

  • Authors:
  • Brendan Elliott;Z. Meral Özsoyoglu

  • Affiliations:
  • Case Western Reserve University, Cleveland, OH, USA;Case Western Reserve University, Cleveland, OH, USA

  • Venue:
  • CIVR '08 Proceedings of the 2008 international conference on Content-based image and video retrieval
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

The number of personal multimedia objects, such as digital photographs and videos, are exploding on the web through popular sites such as Flickr, YouTube, and FaceBook hosting billions of user-created items. Semantic annotation can be an extremely effective way to search, browse, and organize media objects, but can require extensive human involvement. In this work, we show how semantic metadata about social networks and family relationships can be used to improve semantic annotation suggestion. This includes up to 82% recall for people annotations as well as recall improvements of 20-26% in tag annotation recall when no annotation history is available. In addition, utilizing relationships among people while searching can provide at least 28% higher recall and 55% higher precision than keyword search while still being up to 12 times faster. Methods are evaluated on real personal photo collections containing up to 120k photos from Flickr as well as 41k annotated photos from our prototype system.