By Brian Booth – UGent-imec en Dekimo
Description:
In metal 3D printing, roughly 10% of parts are printed with defects so severe that the parts have to be scrapped. With printing processes already optimized, it is believed that further reductions in this scrap rate requires real-time, in-process defect detection and correction. In this talk, we will present our research on high-speed AI modeling for the monitoring of metal 3D printing. In particular, we will highlight how our AI models are designed to detect printing defects and how they run fast enough to enable real-time interventions into the printing process, interventions that improve the quality of the 3D printed parts.
Presenter bio:
Brian Booth is a senior researcher at the imec Image Processing and Interpretation Lab at Ghent University. Prior to joining the lab in 2020, Brian obtained his PhD in computing science at Simon Fraser University (2015) and became a Marie Curie fellow as a postdoctoral researcher at the imec Vision Lab at the University of Antwerp. Brian’s research focuses on the application of AI to real world computer vision problems, particularly problems that are constrained by low time, power, or cost budgets.