Fast and Accurate Recognition of Perovskite Fluorescent Anti-counterfeiting Labels Based on Lightweight Convolutional...

Area of Science:

• Materials Science
• Information Security
• Nanotechnology

Background:

• Anti-counterfeiting is critical for information security.
• Physical Unclonable Function (PUF) labels leverage inherent randomness for security.
• Existing PUF technologies face challenges in scalability and recognition speed.

Purpose of the Study:

• To develop novel, high-security anti-counterfeiting labels using perovskite nanocrystal films.
• To create a high-throughput method for generating controllable perovskite crystal films.
• To introduce an advanced recognition method for enhanced PUF label security.

Main Methods:

• Developed a high-throughput droplet array generation technique using surface tension confinement.
• Constructed PUF textures using the random distribution of perovskite nanocrystal particles.
• Implemented a Partial Convolutional Network (PaCoNet) for PUF recognition.

Main Results:

• Created microscale perovskite PUF labels (< 5.3 × 10⁻² mm²) with fluorescent properties.
• Achieved low cost (< 3 × 10⁻⁴ USD) and high encoding capacity (1.7 × 10¹⁹⁵⁶).
• Demonstrated up to 99.65% recognition accuracy with a recognition time of 0.177 s per image using PaCoNet.

Conclusions:

• Perovskite nanocrystal-based PUF labels offer a promising solution for multilevel anti-counterfeiting.
• The developed fabrication and recognition methods significantly improve security and efficiency.
• These labels have strong potential for widespread application in anti-counterfeiting technologies.