Abstract

In the field of 3D model retrieval, the performance of shape-based method is not satisfactory because of the semantic gap. Therefore semantic-based retrieval is proposed to improve retrieval performance. Ontology theory is an effective means to express semantics, but manual construction of ontology will require tremendous efforts. This paper explored an automatic 3D model semantic annotation and ontology construction method based on semantic relationship among models. Semantic relationship comes from the users' retrieval records of shape-based retrieval and is adopted in the automatic semantic annotation of the models. Experiments on Princeton Shape Benchmark show that the proposed method achieves better performance in both semantic annotation and semantic retrieval.

Paper Details