Solid Model Databases: Techniques and Empirical Results

Abstract
This paper presents techniques for managing solid models in relational database management systems. Our goal is to enable support for traditional database operations (sorting, distance metrics, range queries, nearest neighbors, etc) on large databases of solid models. We introduce an approach to compare models based on shape using information extracted from the model boundary representation into Model Signature Graphs. We show how the Model Signature Graphs can be used to compute topological distances among models and how to use these measures to create metric spaces for indexing and clustering of solid models. We believe this work will begin to bridge the solid modeling and database communities, enabling new paradigms for interrogation of CAD datasets based on the engineering content of solid models.