A Robust Deep Learning Framework for Steganography in 1D and 2D Barcodes

ABSTRACT Barcodes are a key IoT technology for product identification and supply chain traceability, yet they are limited in data capacity and security. To address these limitations, we propose a barcode steganography framework based on image embedding and deep learning. Experiments demonstrate robust information hiding, reliable retrieval and multi‐target recognition under real‐world conditions. Applied to both 1D and 2D barcodes, the method generates image‐augmented codes with significantly enhanced capacity, security and practicality.