Physically Unclonable Fingerprints for Authentication

We have developed an innovative fingerprinting method using the melt-electrospinning printing process for product authentication. This method generates unique, unclonable fingerprints that can be made tamper-proof with a transparent polymer coating. We have successfully tested this approach by printing 393 unique fingerprints on glass substrates, achieving a 95.8% deep learning-based authentication accuracy. Furthermore, fluorescent ink can be employed to enhance fingerprint visibility, enabling analysis through fluorescence microscopy and facilitating spectral authentication. Additionally, the transparent polymer coating obfuscates and encrypts the fingerprint, which can be decrypted using Speeded-Up Robust Features (SURF) techniques. Our ongoing research focuses on assessing the vulnerability of fingerprint images to adversarial attacks, as well as conducting analyses of uniqueness, uniformity, and reliability. We are also ensuring their robustness through machine and deep learning techniques. The proposed authentication scheme aims to provide a dependable solution tailored to the complexities of modern manufacturing and supply chains, effectively mitigating potential intellectual property threats.