Stealth 3D Encryption Based on Spectral Computed Tomography

ABSTRACT Information security is essential across various domains, including daily life, social stability, and national security. However, traditional encryption techniques face insufficient security levels due to two inherent drawbacks: exposed codes and 2D coding. Herein, we propose the concept of stealth 3D encryption (STE) powered by spectral computed tomography (CT). The STE system is constructed using bismuth and barium‐based encoding units, which feature high atomic numbers and different K‐edge values, enabling identification and differentiation by spectral CT. This system not only offers more complex and rich encryption information (up to 2 971 possibilities based on 10 × 10 × 10 encoding units) but also exhibits unique stealth capability, attributable to the high penetration and absolute 3D reconstruction capability of spectral CT imaging. The STE system can be embedded within protected objects and functions effectively even in wooden structures with depths of up to 17 cm, significantly enhancing the protection and security level of the encryption. Additionally, leveraging a U‐net neural network engine, we implement a noise‐robust, comprehensive pipeline for explicit 3D binary pattern design, STE system fabrication, image recognition analysis, and results validation. The proposed STE system opens up a new paradigm for high‐level encryption technology.