Using Multiple Image Representations to Improve the Quality of Content-Based Image Retrieval

Content-based image retrieval (CBIR) has been the object of considerable study since the early 90's. Much effort has gone into characterizing the "content" of an image for the purpose of subsequent retrieval. The present study seeks to capitalize on this work and to extend it by employing content-analysis of multiple representations of an image which we term multiple viewpoints or channels. The idea is to place each image in multiple feature spaces and then effect retrieval by querying each of these spaces and merging the several responses. We show that a simple realization of this strategy can be used to boost the retrieval effectiveness of conventional CBIR.
Note: Abstract extracted from PDF text

French, James, James Watson, Xiangyu Jin, and Worthy Martin. "Using Multiple Image Representations to Improve the Quality of Content-Based Image Retrieval." University of Virginia Dept. of Computer Science Tech Report (2003).