Dual Encryption Technology via Laser‐Induced Perovskite Multicolor and Textures Patterning With Artificial Intelligence Recognition

ABSTRACT Lead halide perovskite films show luminescence ideal for security, but artificial intelligence (AI) advancements have rendered their traditional display encryption obsolete. Herein, for AI recognition, we developed a facile and highly secure dual encryption method based on laser‐induced perovskite multicolor and textures patterning. Precise control of the emission wavelength and surface texture of the perovskite composite film (PCF) was achieved by regulating the composition of the anion exchange solution and laser path. A Quick Response (QR) code consisting of squares with different emission wavelengths and textures was fabricated on CsPbBr 3 PCF. The decryption process requires two keys. The first key is an optical filter, which is used to selectively isolate squares emitting at specific wavelength. Another key is the high resolution surface texture of the patterned squares. Further screening of squares with the specified texture is performed via pattern recognition using the YOLOv5 model, thereby achieving decryption of the QR code content. The YOLOv5 model achieved an average similarity of 89.5% in recognizing training samples, with a precision of 96.44%, which means the chosen model exhibited effective recognition performance. This strategy successfully realized dual encryption of information and provided a novel solution for information security protection with AI recognition.