Loading...

Reducing Semantic Gap in Content-Based Image Retrieval Systems Using Graph Cuts and Fuzzy Relevance Feedback

Shafeian, Hessamoddin | 2009

603 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 39895 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Tabandeh, Mahmoud
  7. Abstract:
  8. Multimedia retrieval systems are gradually playing a critical role in our everyday life to facilitate interacting with massive amount of personal or professional images, music and video archives. So far, many systems have been proposed among them relevance feedback based content based multimedia (especially image) retrievals has been proved to be more effective. However there is still a problem called semantic gap, in finding proper mapping between low-level features used by CBIR systems and user’s high-level concepts. On the other hand graph cuts have been a great powerful tool for solving many computer vision problems. They benefit from robust optimization algorithm called maximum flow/ minimum cut. In this thesis, we combine their power with advantages of a new kind of image features called modified virtual features (MVF)[17]. The image library is considered as an undirected graph, whose nodes are suggesting the images itself and edges are representing a proximity relation between image pairs. After running maximum flow/ minimum cut module, image database is divided into two sets of relevant and irrelevant images to user information need. This cut acts as a filtering tool to discard irrelevant images and bridging the semantic gap. As a result, images database shrinks to a smaller set of closer images to the query, which in turn results in a better retrieval accuracy and reduced runtime. We also investigate the effect of considering fuzzy judgment of the user which is a more natural form of decision making. Experimental results prove the effectiveness of our contributions.
  9. Keywords:
  10. Image Retrieval ; Maximum Flow ; Fuzzy Membership Function ; Content Based Retrieval ; Virtual Barrier ; Graph Cut

 Digital Object List

 Bookmark

No TOC