Digital Engineering 24/7 : Purdue University Uses Machine Learning to Classify Mechanical Objects

by | Nov 23, 2020

Professor Karthik Ramani was interviewed by Kenneth Wong of Digital Engineering 24/7 regarding the recent release of the Mechanical Component Benchmark dataset.

You can listen to Professor Ramani’s interview on Digital Engineering’s web site:

In early 2000, Dr. Karthik Ramani experimented with a shape-based search system that allows engineers to use visual references to identify and search for mechanical objects, but it soon became clear to him the computing power was insufficient.

Twenty years later, with far more computing power at his disposal, he revived the project, applying machine learning (ML) to crawl through a database of gearboxes, bearings, brakes, clutches, motors, nuts, bolts, and washers. The outcome is an open-source annotated database of more than 58,000 3D mechanical parts.