Chinese Journal of Computers   Full Text
Title	A Survey of Content-Based 3D Model Retrieval with Semantic Features
Authors	PAN Xiang1) ZHANG San-Yuan2) YE Xiu-Zi2)
Address	1)(Institute of Computer Science, Zhejiang University of Technology, Hangzhou 310014) 2)(Institute of Computer Science, Zhejiang University, Hangzhou 310027)
Year	2009
Issue	No.6(1069—1079)
Abstract &   Background	Abstract Recent researches shows shape features is not enough for Content-based 3D retrieval. More and more researches are focusing on 3D retrieval by semantic features. This paper gives a full review of recent work about semantic-based 3D retrieval. Firstly, it makes an analysis about the existing researches, including extraction of semantic feature, active learning for semantic retrieval and semantic related retrieving platform. Secondly, this paper gives a comparison of some typical researches and concludes their contributions as well as limitations. Finally, some challenging research topics are proposed as a future work. Keywords content-based 3D retrieval; semantic features; design reusing; ontology; active learning Background This project is supported by the National Natural Science Foundation of China under Grant “Hierarchical Structure Semantic Extraction for 3D Model Retrieval” (No.60703001). The project is mainly to propose a framework in extracting structural semantic feature. For 3D models, this framework can get high-level semantic knowledge representation automatically from low-level geometry information. It is the most important problem in 3D model retrieval and reusing. Currently, there are lots of review papers for 3D model retrieval. However, these papers only discuss shape feature extraction and similarity. Few are focusing on semantic based 3D retrieval. Semantic based 3D retrieval has attracted more and more research interests. It achieves some research achievements but remains many unresolved problems. Therefore, this paper gives a full review about semantic based 3D retrieval and recommends some challenging research topics for future works. It can offer a summarized reference to interested researchers.