Hierarchical Layout Deduction for Furniture Model Retrieval

3D CAD model retrieval plays a fundamental role in producing impressive and innovative designs easier and faster. However, after decades of efforts, effective retrieval approach for furniture models is still rare. A novel retrieval approach for furniture models, based on a hierarchical layout deduction, is proposed according to the common structure characteristics of furniture. In order to discriminate furniture models effectively both on topological, geometrical and semantic layers, special components for furniture model are defined and a multiresolution furniture model descriptor, attributed hierarchy layout graph, is adopted. To promote the retrieval efficiency, a filtering process, based on random walks method, is applied. Based on hierarchy graph matching, the similarity between two furniture models, is determined. A prototype system is implemented to verify the proposed approach effectiveness.