A review and assessment of 3D CAD model retrieval in machine-part design

With the widespread adoption of digital modeling and manufacturing technologies, the volume of 3D CAD model data is rapidly increasing. Reusing existing resources can shorten new product development cycles, reduce development costs, and improve the performance and quality of new products. This paper provides an overview of the advancements in 3D model retrieval technology within the design process. Based on the dimensions of query models, the research progress in 3D CAD model retrieval is systematically summarized and discussed. This encompasses methods based on 3D models (including statistical retrieval, geometric shape analysis, and topology analysis methods), 2D images (such as view mapping and multiview projection methods), and semantic technologies (including ontology retrieval and knowledge graph retrieval methods), all considered in light of design stage requirements. This review outlines the challenges associated with retrieving 3D CAD models in part design, discusses the principles underlying existing retrieval methods, reviews recent research progress in this area, systematically examines the strengths and weaknesses of each method, and suggests future research and development directions.