Contextual Image Search

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

Bibtex

@inproceedings{cis:mm:2011,
 author = "Wenhao Lu, Jingdong Wang, Xian-Sheng Hua, Shengjin Wang, Shipeng Li",
 title  = "Contextual Image Search",
 booktitle = "ACM Multimedia",
 year   = "2011"
 }

Results

Image search with textual contexts   Example 1 Image search with the masked query "Micheal Jordan" Image search with query "Micheal Jordan" and textual context "dunk" and "slam"   Example 2 Image search with the masked query "Cambridge England" Image search with query "Cambridge England" and textual context "river", "boat" and "floating"   Example 3 Image search with the masked query "George Bush" Image search with query "George Bush" and textual context "joke", "fool" and "prank"   Image search with visual contexts   Example 4 Image search with the masked query "Queen Victoria" Visual context Image search with query "Queen Victoria" and visual context   Example 5 Image search with the masked query "tower bridge" Visual context Image search with query "tower bridge" and visual context