Title :
Refining image retrieval using one-class classification
Author :
Xiao, Jie ; Fu, Yun ; Lu, Yijuan ; Tian, Qi
Author_Institution :
Dept. of Comput. Sci., Univ. of Texas at San Antonio, San Antonio, TX, USA
fDate :
June 28 2009-July 3 2009
Abstract :
Can we take advantage of the huge number of online images to improve image search quality? Motivated by this question, we propose a novel model to re-rank Google image search results by exploring the latent characteristic of massive unrelated images as a clue to filter them in the reranking. Inspired by the characteristic of the intrinsic diversity and the unwanted availability of the unrelated images, in our model, we adopt one-class classification to build a hyper-sphere for the target objects, unrelated images, and construct a robust boundary to distinguish them from the related images effectively. Then the Google results can be easily re-ranked by filtering the unrelated images with the built-up model. Extensive experiments demonstrate our approach outperforms Google image search engine´s results, even if its baseline is high.
Keywords :
image classification; image retrieval; information filtering; search engines; Google image search quality; hyper-sphere classification; image filtering; image re-ranking; image search engine; intrinsic diversity; latent characteristics; one-class classification; online image retrieval; target objects; Computer science; Feedback; Filtering; Filters; Image representation; Image retrieval; Keyword search; Robustness; Search engines; Web search; One-class classification; image retrieval; re-ranking; relevance feedback;
Conference_Titel :
Multimedia and Expo, 2009. ICME 2009. IEEE International Conference on
Conference_Location :
New York, NY
Print_ISBN :
978-1-4244-4290-4
Electronic_ISBN :
1945-7871
DOI :
10.1109/ICME.2009.5202498