Abstract: Counterfeit components result in billions of dollars of financial loss every year. Counterfeit components also impact product reliability and stability: this creates loss of confidence in consumers and, for mission critical components, can result in catastrophic system failures. Component traceability is an important aspect of ensuring that counterfeit components are kept out of the supply chain. Several researchers have attempted to use 2D imaging to record intrinsic part features that can be used to uniquely identify parts. This approach works for situations where the surface of the part does not change. 2D imaging is also not reliable for situations in which the camera pose, intrinsics, and other critical parameters (e.g. sensitivity, resolution, ambient light) change between the acquisition camera and the verification camera. This talk will focus on the use of advanced 3D imaging methods combined with convolutional neural networks that overcome the limitations of 2D imaging to effectively track parts throughout their operational life. This includes situations where the surface of the part changes.

Bio:Matt Bellis is the founder and CEO of Seikowave, a company focused on high-speed, microscopic 3D imaging. Prior to founding Seikowave, Mr. Bellis was part of the executive team at two separate high-tech startups focused on the development of micro-electromechanical (MEMS) devices. Both MEMS companies were acquired. Earlier in his career, Mr. Bellis worked for Northrop Grumman and Hewlett-Packard where he held positions in R&D, Marketing and Management. Mr. Bellis has patents in disk drive servo-track writing, projection displays, MEMS device design, and 3D imaging. He holds Bachelor degrees in Electrical Engineering, Physics, and Material Science from Northwestern University and an MSEE from UC California at Santa Barbara where he concentrated on compound semiconductor lasers. Mr. Bellis lived in Japan for ten years and is proficient in Japanese.