A facile way to construct highly stable PUF tags for unclonable anti-counterfeiting and authentication with computer vision

Abstract

Constructing a robust and physically unclonable anti-counterfeiting label with a facile but effective concept remains challenging these days. In this work, we proposed a new and facile strategy by integrating phase separation induced polydimethylsiloxane (PDMS) patterns with computer vision technique to construct an anti-counterfeiting system. PDMS patterns embedded with randomly distributed bubbles were devised by simply spraying water droplets onto the surface of uncured PDMS coatings. With the aid of computer vision algorithms, the patterns can be encoded as standard labels and stored in a library. Subsequently, the standard labels can be used to identify the unknown patterns by decoding algorithm. The good properties of the PDMS patterns including high randomness and superior stability against harsh environments, as well as the excellent performance of the computer vision algorithm were all confirmed in this study. To the best of our knowledge, this is the first time to create physically unclonable anti-counterfeiting labels by utilizing the phase separation effect between water and PDMS. Moreover, this new strategy has several advantages over the existed strategies such as fluorescent labels and structural color labels. This work would provide a new solution to create security labels.