Toward Effective Mechanical Design Reuse: CAD Model Retrieval Based on General and Partial Shapes

Abstract

In product design, a large proportion of three-dimensional (3D) computer-aided design (CAD) models can be reused to facilitate future product development due to their similarities in function and shape. This paper presents a novel method that incorporates modeling knowledge into CAD model similarity assessment to improve the effectiveness of reuse-oriented retrieval. First, knowledge extraction is performed on archived feature-based CAD models to construct feature dependency directed acyclic graph (FDAG). Second, based on the FDAG subgraph decomposition, two useful component partitioning approaches are developed to extract simplified essential shapes and meaningful subparts from CAD models. Third, the extracted shapes and their FDAG subgraphs are indexed. Finally, the indexed shapes that are similar to user-sketched queries are retrieved to reuse, and FDAG information of the retrieved shapes is provided as redesign suggestions. Experimental results suggest that the incorporation of modeling knowledge greatly facilitates CAD model retrieval and reuse. Algorithm evaluations also show the presented method outperforms other 3D retrieval methods.