Collaborative Research: Facilitating Supply Chain Trust via Micro-Surface Sensing and Vision-Enabled Authentication

Counterfeit products prevail due to untrusted supply chains and have been inflicting significant harm to our society at all scales, from public health, to the economy, and to national security. Distributed ledger technologies such as blockchain have recently been applied to provide unforgeable digital records and to improve the trust of the global supply chains. Despite the improved trust in cyberspace, duplication attacks in the physical world are still possible, making counterfeits remain a serious concern, especially for products of high value. Commonly used QR codes or low-cost RFID tags do not prevent counterfeiters from linking fake goods to the distributed-ledger-based supply chain in cyberspace, which compromises the trust of the overall supply chain. To address this vulnerability, we propose establishing a secure cyber-physical link using physically unclonable features (PUFs) of an item’s surface at the microscopic scale. These microscopic features constitute a “fingerprint,” which can uniquely characterize and identify the specific item being protected and can be captured by consumer-grade cameras. This research project focuses on microsurface sensing and vision-based authentication and offers a novel solution to complement the existing blockchain-based countermeasures in supply chain management. This non-forgeable cyber-physical link provides proactive authentication by supply chain participants/end-users. Successful research in this area can mitigate the illicit reuse of old electronic components and prevent counterfeit medicines from reaching patients.The proposed research effort aims at investigating the scientific foundation and overcoming technological barriers that utilize surface PUFs for linking the physical world to supply chains securely and uniquely via ubiquitous mobile imaging. First, surface PUFs in mobile imaging settings are examined to uniquely identify objects, with a focus on designing efficient and scalable computer vision algorithms. Second, microscopic-level light reflection models are established via physics-guided signal analysis and machine learning approaches to extract surface PUF signals. Third, the integration of the surface PUF verification with the blockchain-based supply-chain information management system is explored and multiscale data collections are planned. The proposed research is expected to advance the scientific understanding of photometric behaviors of microstructures and result in more accurate light reflection models at the microscopic scale via both controlled experiments and machine learning. The information-theoretic modeling of the PUF’s inherent randomness and channel randomness due to processing steps can potentially help identify the bottleneck of the verification system and improve its performance. The proposed ambient light decomposition and estimation scheme facilitate a capability leaping of computer vision for authenticating scenes at the microscopic scale. The proposed research program has the potential to protect public health and national security via safeguarding supply chains. The ubiquity of mobile devices multiplies the economic and societal impacts of our research. The proposed program also seamlessly integrates research, education, and outreach. It will prepare the future security-aware workforce by fascinating them with cutting-edge research results via lectures, projects, and global competitions.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.