Optical Labeling with Artificial Intelligence Using Infrared-Responsive Functional Textiles

Functional textiles integrated with image recognition systems hold significant potential for advanced applications. However, conventional visible light dyes and colorants can be detected under normal lighting conditions and are difficult to conceal within materials. In this study, we introduce a sustainable and intelligent textile platform by integrating SiO₂/Al₂O₃-coated hydrophilic Ag₂S quantum dots into natural pineapple leaf fiber (PALF). This biogenic composite exhibits excellent photostability and water compatibility, with concealed optical functionality detectable only under short-wave infrared imaging. Unlike conventional visible dyes, the embedded emission remains invisible under ambient conditions, enabling covert information encoding. To achieve material-level recognition, a convolutional neural network (CNN) was trained on hyperspectral imaging data and successfully identified optically functionalized fibers with 82% accuracy. This was demonstrated even in millimeter-scale yarn samples (∼0.6 cm). This work exemplifies the synergy between renewable materials, nanophotonics, and artificial intelligence, offering practical utility in secure labeling, anticounterfeiting, and interactive textiles.