Exploit camera metadata for enhancing interesting region detection and photo retrieval

  • Authors:
  • Zhong Li;Jianping Fan

  • Affiliations:
  • Department of Computer Science, The University of North Carolina at Charlotte, Charlotte, USA 28223-0001;Department of Computer Science, The University of North Carolina at Charlotte, Charlotte, USA 28223-0001

  • Venue:
  • Multimedia Tools and Applications
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Photographs taken by human beings differ from the images that taken by a lifeless device, such as a surveillance camera or a visual sensor on a robot, in that human being intentionally shoot photographs to express his/her feeling or photo-realistically record a scene. This creation process is accomplished by adjusting two factors: the setting of parameters on a camera and the position between the camera and the object which he or she is interested in. In this paper, this procedure is learned using the machine learning technique so that what the interest of the photographer is and what the core content of a photo wants to display can be reversely calculated. A photo retrieval system was built upon the category of interesting regions and the metadata is used for help. The research also explore the argument how local or global feature affects the performance of image retrieval. A novel stochastic segmentation algorithm called region restricted EM algorithm was applied in order to construct the interesting regions. Experimental evaluation on over 7000+ photos taken by 200+ different models of cameras with variety of interests has shown the robustness of our technique.