Abstract: 

The shortage in electronic components caused an increase in the procurement of components from unauthorized distributors and a spike in the occurrence of counterfeits in the market. Conventional methods to mitigate this risk by performing lab inspection on samples of components within a batch according to the AS6171 standard or a more advanced 100% inspection using in-line visual authentication using the vision system in place in the SMT line.

However, the AS6171 tests are performed only on the material procured in the free market leaving all the other material with no verification. In addition, the in-line inspection takes place during the operation of the line preventing the material from being rejected during acceptance. In this work, we present a simple screening method that allows the filtering materials as they are accepted into storage by verifying the authenticity of the labels on the components containers.

The labels are easily copied or reproduced; nevertheless, they contain data that can be used to cast suspicion on a container and to flag it for deeper investigation. The authentication method presented is built on layers of algorithms addressing different authentication aspects. The first layer uses deep networks to verify the label's visual features in reference to the label's historical data scanned for the same apparent source. An ML classifier detects the similarity of all the visual features like font, boldness, roundness, logos, and distribution in the label. A second layer verifies the correlation between the barcodes and the text layer on the label. A third layer
compares the relation between the lot numbers and the date codes based on previously accumulated data in the database. And a final layer that integrates the results of the lower layer into a recommendation.

While this method by itself does not address all counterfeit cases, it can dramatically reduce the risk of using counterfeit materials in the acceptance stage as it can be easily applied to all accepted materials. In addition, all the logistical information obtained from the labels can be further used to automatically accept the materials into storage without manual data entry.

Biography:

Dr. Eyal Weiss [Cybord]

Dr. Eyal Weiss specializes in multidisciplinary technology development. Awards and accolades winner in the fields of machine learning, plasma physics, optical assemblies, laser technology, and electromagnetics, including twice the prestigious “Israel security prize”. Dr. Weiss lead innovative research at Soreq Research Center (SRC) for accumulated 17 years in various complex and multidisciplinary technological fields combining cutting-edge technology with extremely advanced AI. He has designed and built numerous manufacturing production lines utilizing new and disruptive technology.  Dr. Weiss holds B.Sc in Mechanical Engineering and M.Sc. in Plasma Physics from the Technion-Israel Institute of Technology, and a Ph.D. in Electronic and Computer Engineering from the Ben-Gurion University of the Negev. In 2018, he founded and became CTO of Cybord, developing electronic component qualification and authentication technologies. He is a member of the Israel Innovation Authority, Euromet, SAE, and IPC committees. Dr. Weiss is an expert in technology development and manufacturing technology and had published over 25 peer-review articles, 4 patents, and a book.