Abstract

Due to 3D CAD models are often characterizedwith complicated geometry and topology, how to help thedesigners quickly and accurately find the object modelsfrom the database containing massive amount of modelsusing the relatively rough query instance is still a bigchallenge. In this paper, a novel representation of 3D CADmodels using hierarchical graph (HG) is proposed. Themodel descriptors are divided into shape feature descriptorsand topology relationship descriptors. These two descriptorscan be extracted from HG. In this way, coarse-grained andfine-grained 3D CAD model retrieval are implemented. Onthis basis, a 3D model retrieval method is proposed based ongenetic algorithm (GA) and ant colony optimization (ACO),which are employed to detect the common sub-graph in thecorresponding the hierarchical graphs of different models.This method can improve the accuracy and efficiency of 3DCAD model retrieval. Based on the above researches, a 3Dmodel retrieval system HUST-CMRS is developed. Ourexperimental results show that, the algorithm proposed inthis paper can implement multi-mode indexing and satisfypersonalized need of users.