High-dimensional anticounterfeiting nanodiamonds authenticated with deep metric learning.

Abstract in brief

• Physical unclonable function (PUF) labels are seen as a strong solution for preventing counterfeiting, but face challenges such as limited high-quality materials and ineffective authentication methods.
• The authors introduce a new method using fluorescent nanodiamonds that allows for three-dimensional encoding, achieving high-quality digitized images quickly and with great stability.
• They also propose a deep metric learning algorithm for authentication, which improves noise tolerance and is effective even with previously unseen labels, making it a more reliable option than traditional methods.