Automated Counterfeit Detection
Verify product authenticity of electronic products returned for refund
Graphical programming environment for deep learning-based industrial image analysis
High-end smartphones and other expensive consumer electronic products are natural targets for counterfeiting. Counterfeit product returns for refund is an increasing problem and source of financial losses for manufacturers.
Products are purchased and returned via e-commerce. Many ostensible product returns are actually counterfeits that use the same or similar case, but contain other internals, or are empty. Manufacturers must verify product authenticity before issuing a refund.
X-ray images can show the interior of each case. Given the wide range of possible contents, conventional machine vision is limited in its ability to distinguish authentic from counterfeit products at the rate needed for effective processing.
Cognex Deep Learning is an ideal solution for identifying fake product returns since it does not rely on complex and time-consuming programming. The part location tool trains on a set of images containing the specific parts that need to be present for the returned smartphone to be classified as complete and authentic. Once trained, the part locate tool examines X-ray images of returned phones and quickly confirms the presence or absence of these parts in the proper locations without any need to open the case.
Depending on the types of counterfeits being returned, Cognex Deep Learning’s classification tool might also be used. In this case, the classification tool trains on X-ray images of both authentic and counterfeit phones and learns to distinguish authentic phones from a wide range of counterfeits. As counterfeit types change, the classification tool easily retrains on a set of images of the new type and can be quickly re-deployed on the line with the new category, without programming.