Content area
Full Text
Knowl Inf Syst (2009) 21:113132
DOI 10.1007/s10115-008-0183-x
REGULAR PAPER
Zhiguo Gong Qian Liu
Received: 16 April 2008 / Revised: 15 October 2008 / Accepted: 8 November 2008 / Published online: 12 December 2008 Springer-Verlag London Limited 2008
Abstract This paper discusses techniques for improving the performance of keyword-based web image queries. Firstly, a web page is segmented into several text blocks based on semantic cohesion. The text blocks which contain web images are taken as the associated texts of corresponding images and TF*IDF model is initially used to index those web images. Then, for each keyword, both relevant web image set and irrelevant web image set are selected according to their TF*IDF values. And visual feature distributions of both positive image and negative image are modeled using Gaussian Mixture Model. An images relevance to the keyword with respect to visual feature is thus dened as the ratio of positive distribution density over negative distribution density. We combine the text-based relevance model with visual feature relevance model to improve the performance. Thirdly, a query expansion model is used to improve the performance further. Expansion terms are selected according to their cooccurrences with the query terms in the top-relevant set of the original query. Our experiments show that our approach yield significant improvement over the traditional keyword based query model.
Keywords Web image Web search Associated text Visual feature Query expansion
1 Introduction
With the explosive growth of the web, millions of images, on almost any subject, are available on the web. How to effectively and efciently reuse this comprehensive and huge image resource is drawing more and more attention from both academic and industrial communities. Many web image search engines, either as commercial systems such as Google image search (http://images.google.com/
Web End =http://images.google.com/ ), Yahoo Image Search (http://images.search.yahoo.com
Web End =http://images.search.yahoo.com ), Lycos (http://lycos.com/
Web End =http://lycos.com/ ), AltaVista photo nder (http://www.altavista.com/
Web End =http://www.altavista.com/ ), Ditto
Z. Gong (B) Q. Liu
Faculty of Science and Technology, University of Macau, Macao, Peoples Republic of China e-mail: [email protected]
Q. Liue-mail: [email protected]
Improving keyword based web image search with visual feature distribution and term expansion
123
114 Z. Gong, Q. Liu
(http://www.ditto.com/
Web End =http://www.ditto.com/ ), art.com (http://www.art.com
Web End =http://www.art.com ), Amico (http://www.amio.org
Web End =http://www.amio.org ),...