Application of discriminant-EM in image retrieval by relevance feedback

Document Type

Conference Proceeding

Publication Date

12-1-2004

Abstract

In this paper, we introduce a new method for retrieving images in an image database. In previously proposed Relevance Feedback methods, the user is asked to label the image examples as positive (relevant) or negative (irrelevant) images. In our model, the user can label the example images with a score between zero and one. We implemented an Expectation Maximization (EM) algorithm to be able to use unlabeled data as well as labeled data for clustering images. Also, we use discriminant analysis to give higher weights to features that are more important in discriminating relevant from irrelevant images.

Publication Source (Journal or Book title)

IIE Annual Conference and Exhibition 2004

First Page

2223

Last Page

2228

This document is currently not available here.

Share

COinS