Image Search by Concept Map

Hao Xu1, Jingdong Wang2, Xian-Sheng Hua2, and Shipeng Li2
1University of Science and Technology of China  2Microsoft Research Asia
 

 

Abstract

In this paper, we present a novel image search system, image search by concept map. This system enables users to indicate not only what semantic concepts are expected to appear but also how these concepts are spatially distributed in the desired images. To this end, we propose a new image search interface to enable users to formulate a query, called concept map, by intuitively typing textual queries in a blank canvas to indicate the desired spatial positions of the concepts. In the ranking process, by interpreting each textual concept as a set of representative visual instances, the concept map query is translated into a visual instance map, which is then used to evaluate the relevance of the image in the database. Experimental results demonstrate the effectiveness of the proposed system.

 

Download

| PDF |
 

Bibtex

@inproceedings{ConceptMap:2010,
  author = "Hao Xu and Jingdong Wang and Xian-Sheng Hua and Shipeng Li",
  title  = "Image Search by Concept Map",
  booktitle = "SIGIR",
  year   = "2010"
}
 

Results

©Copyright Jingdong Wang 2010
HTML Hit Counters