Additive Manufacturing Magazine showcases Holm's computer vision research

MSE’s Elizabeth Holm and her research team developed a computer vision system that characterizes metal powders used in additive manufacturing (AM) with up to 95 percent accuracy. Where humans only achieve 50 percent accuracy, Holm’s system uses micrograph images to identify, characterize, and ‘fingerprint’ powders based on qualitative and quantitative properties that human experts simply can’t classify. Holm’s goal is bigger than teaching computers to identify powders. She sees her system as a path to utilizing the data inherently produced by AM processes, allowing for the identification of material changes and maximization of print quality. “In additive manufacturing, we’re in a situation where, by the nature of the process itself, we are going to be given a lot of data. What machine learning is really good at is taking data and making some sense of it—finding correlation and trends and directions,” Holm told Additive Manufacturing. read more