A search mechanism for identifying and retrieving 3D CAD components from a repository based on 2D/3D shape inputs.

Allow search and retrieval from a database of 3D CAD components based on geometric 3D geometry or 2D sketches as inputs. Develop the solution for easy integration with existing CAD tools, Product Lifecycle Management Solution and as a base for building shape based engineering applications. Develop based on standard geometry representation such as VRML for easy portability.

Engineering Industry captures a large volume of design knowledge in the form of 3D CAD models. Typical NPD or product modification involves engineers looking for existence of similar components and alternate designs from this repository. Initiatives in Lean Engineering targets part count reduction, common parts, variant creation etc and require accurate and timely identification of such components. Today, widely adopted methods are limited to attribute based metadata search and classification. All of these require manual data input and determination of the class. Due to the ever changing nature of the industry and classification procedures, the efficiency decreases over time and cost of maintaining a usable knowledgebase of CAD components remains high. Geometry or shape of a component is an invariant element that can be searched for and retrieved regardless of the manual classifications or application.

3 solution providers have commercialised shape search, however in limited integrated form. A sample test conducted on the tools exposed considerable gaps in the accuracy, integration and usefulness of the solution compared to expectations.
The challenge

Comparing large databases of complex CAD geometry & classification involve complex computational geometry techniques and results on accuracy and speed of available methods have so far given varying degree of confidence.

Advancement in computational geometry and applications in the field of computer vision has provided many methods for visual shape descriptor creation and matching techniques that can be successfully applied to engineering geometry. Supported with a classification method purely based on shapes, the complexity of the search problem can be brought to manageable levels with sufficient accuracy.

Sconce released its preimier high end Shape Search Product - Bingo! at the PTC User event on June 9, 2009 at Orlando, Florida, USA. Driven by unique, powerful and accurate shape identification engine, Bingo! enables automatic part classification, efficient indexing based on shape, and fast similar-search on small to large library of 3D models. Bingo! is designed to integrate as an essential tool for companies adopting Design Reuse. Design Reuse is extremely important and survey says that 30% to 40% of new parts designed either already exists or have similar parts that can be used as a base model to start.Bingo!'s CAD/PLM neutral architecture integrates easily with all the major CAD and PLM systems.