Multilevel Relevance Feedback for 3D Shape Retrieval

Abstract

Relevance feedback techniques are expected to play an important role in 3D search engines, as they help to bridge the semantic gap between the user and the system: similarity is a cognitive process, depending on the observer. We propose a novel relevance feedback technique, whose basic idea is threefold. First of all, the user is provided with a variety of shape descriptors, analysing different shape properties. The user then expresses her similarity concept through a friendly interface which supports multilevel relevance judgements. Finally, the system inhibits the role of the shape properties that do not reflect the user's idea of similarity. The assumption is that similarity may emerge as an inhibition of differences, i.e., as a lack of diversity with respect to the shape properties taken into account. The proposed technique is based on a simple scaling procedure, which does not require any a priori learning or optimization of parameters.