Relevance feedback for content-based image retrieval using the Choquet integral

TLDR: A more sophisticated model for similarity judgments based on fuzzy measures and the Choquet integral is explored, and a suitable algorithm for relevance feedback is proposed that is preferable to traditional weighted-average techniques.

Abstract: Relevance feedback is a technique to learn the user's subjective perception of similarity between images, and has recently gained attention in content based image retrieval (CBIR). Most relevance feedback methods assume that the individual features that are used in similarity judgments do not interact with each other. However, this assumption severely limits the types of similarity judgments that can be modeled. The authors explore a more sophisticated model for similarity judgments based on fuzzy measures and the Choquet integral, and propose a suitable algorithm for relevance feedback. Experimental results show that the proposed method is preferable to traditional weighted-average techniques. The proposed algorithm is being incorporated into a CBIR system developed at Korea Telecom.