Adaptive Similarity Measurement Using Relevance Feedback
Abstract
Content-based image retrieval (CBIR) is the core technology for many applications. Many researchers have interested in how to extract the important features in the image for the CBIR. However, different applications have their own emphasized image features. In this paper, we proposed a novel customized relevance feedback (RF) mechanism which can set adaptive weights of similarity measurement for each database image from the user feedback. Through this mechanism, we could analyze customized retrieval habit and standpoint to gauge proper features to adjust similarity measurement. System can improve the retrieval precision/recall, and make each user satisfied with retrieval results. Moreover, the experiments present improved ratio of precision (or recall) is notable.
Related Papers
- → Evaluating group-based relevance feedback for content-based image retrieval(2004)26 cited
- → Semantic image retrieval using relevance feedback and reinforcement learning algorithm(2010)3 cited
- → Query-by-sketch image retrieval using relevance feedback(2005)4 cited
- → A Relevance Feedback Image Retrieval Approach Based on RGA(2008)
- RESEARCH ON MULTI-LEVEL LOG-BASED RELEVANCE FEEDBACK SCHEME FOR IMAGE RETRIEVAL(2012)