Deep Learning accelerates X-ray CT inspection of 3D-printed parts

Deep learning improves the speed and accuracy of X-ray CT inspections for 3D-printed components
The Department of Energy’s Oak Ridge National Laboratory has developed a new deep-learning framework that speeds up the inspection of metal parts made using additive manufacturing, and increases the accuracy of results. Reduced costs in time, labor and maintenance, as well as energy, are expected to increase the use of 3D printing or additive manufacturing.

Amir Ziabari, ORNL’s lead researcher, said that the scan speed reduced costs by a significant amount. The quality of the scans is better, which makes the analysis much easier.

The framework has already been incorporated into the software that ZEISS uses in its machines at DOE’s Manufacturing Demonstration Facility, ORNL. This facility is where companies hone their 3D printing methods.