A point-subset compression algorithm for fiber-based certificates of authenticity

This paper discusses the certificates of authenticity (COAs) that are digitally signed physical objects with a unique random structure. The random structure of a fiber-based COA relies on the fact that if one end-point of a fiber is exposed to light, the other one illuminates. A COA instance is compressed and then combined with the cryptographic hash. Each COA instance is associated with message recovery using issuer's public key. The key optimization goal in this system is to compress as much of the entropy of COA's random structure. The key parameter of the point-compression problem is: computing the COA model, encoding pixel-to-pixel vectors, and heuristically solving the asymmetric traveling salesman problem. The encoding can be done using arithmetic vector. A heuristic is developed that aims at solving point-compression problem and the compression rate is obtained much better.