关键词: 三维CAD模型, 检索, 局部特征, 图谱, 空间词袋, 层次特征描述子

Abstract: To improve the distinguishing ability for model's local parts and the searching efficiency in 3D CAD model retrieval,a novel algorithm using hierarchical feature descriptor was proposed based on graph spectrum theory and spatial bag of words model.By extracting B-Rep information such as geometry and topology attributes of the models,3D CAD models were represented by Attribute Adjacency Labelled Graphs (AALGs).The models in training data set were decomposed into surface regions with different convexity,which were described as local feature vectors using graph spectrum,and the codebook was created with clustering algorithm.The hierarchical feature descriptor of each library model was constructed,according to which a two-level searching mechanism was employed in CAD model retrieval.Experimental results showed that the proposed method could improve both retrieval performance and efficiency greatly,which achieved the effective reuse of CAD model.

Key words: 3D computer aided design model, model retrieval, local features, graph spectrum, spatial bag of words, hierarchical feature descriptor