Research on 3D CAD Model Retrieval Algorithm Based on Global and Local Similarity

Abstract

Content-based 3D CAD model retrieval takes a 3D CAD model as input and finds other models with the same or similar structure. This paper proposes a two-stage retrieval method that can take into account the global and local similarity of CAD models. In the first stage, the CAD model formation candidate modles with high global matching degree with the query model is selected, and the TF-IMF (Term Frequency-Inverse Model Frequency) vector method is proposed to describe the global surface line distribution of the 3D CAD model. In the second stage, on the basis of the global similarity, the models which are locally similar with the query models are further filtered, and the attribute adjacency graphs between models are calculated by ACO (ant colony optimization) algorithm. Experimental results show that the proposed method achieves better retrieval performance than the maximum clique method based on the attribute adjacency graph (NDCG), which is 90.68%, and has higher retrieval efficiency.