Retrieving Keyword's to an Image Query using Kernel CCA
David R. Hardoon, Sandor Szedmak and John Shawe-Taylor
(2004) Working Paper. University of Southampton, Southampton, UK.

Abstract

In this paper we propose an approach to automatically annotate queryimages with keywords. We use kernel Canonical Correlation Analysis to learn a semantic representation between images and their associateddocuments. The semantic space provides a common representation and enables a comparison between the documents and images. This represen-tation is then used in the creation of new document, comprised from the keywords that best fit the image query. We compare our method againsta standard cross-representation retrieval technique known as Generalised Vector Space Model.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.

[Edit]