Space Transformation Based Approach for Effective Content-Based Image Retrieval

In this paper, we extend the work done by Choubey and Raghavan, which proposed an approach to content-based image retrieval that uses the space transformation methods proposed by Goldfarb to transform the original low-level image space into a feature space, where images are represented as vectors of high-level features that enable efficient query processing. The resulting retrieval system consists of three phases: database population, online addition and image retrieval. In the current work, we investigate issues relating to online addition of new images, during database population phase, in incremental stages. The results show that our approach is effective in retrieving images even when the image database is dynamic.

• Type: Conference
• Year: 2003
• Where: ISMIS
• Authors: Biren Shah, Vijay V. Raghavan
• Keywords: Content-based Image Retrieval, Image Retrieval, ISMIS 2003, Low-level Image Space