Optical Labeling with Artificial Intelligence Using Infrared-Responsive Functional Textiles

TL;DR

This paper introduces a smart textile that uses hidden infrared-responsive materials and AI to encode and recognize information covertly.

Contribution

A novel biogenic composite textile with concealed optical functionality and AI-based recognition for secure labeling.

Findings

A SiO2/Al2O3-coated Ag2S quantum dot composite was developed with infrared-responsive optical properties.

A CNN achieved 82% accuracy in identifying optically functionalized fibers using hyperspectral imaging.

The system enables covert information encoding in textiles under ambient conditions.

Abstract

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.

Figures