ABSTRACT The digitalization of ceramic technology teaching generates sensitive multi‐modal data, including personal information and valuable cultural assets like unique ceramic patterns, requiring robust protection. This paper establishes a comprehensive, network security‐empowered algorithm system for such teaching scenarios. The system integrates multi‐modal sensitivity identification, feature fusion, hybrid encryption, and a reversible image perturbation mechanism. Key results show the sensitivity identification model achieves 0.97 accuracy. The hybrid encryption keeps transmission delay below 17.5 ms for real‐time teaching. The reversible perturbation algorithm effectively protects cultural images (e.g., Fanchang kiln celadon), with authorized restoration SSIM (Structural Similarity Index Measure) at 0.92–0.95 and unauthorized restoration PSNR (Peak Signal‐to‐Noise Ratio) reduced to 18–20 dB, hindering pattern theft. The main contribution is a unified security framework that maintains controllable risk for teaching and cultural data throughout the digital pipeline, providing a deployable foundation for the secure digital inheritance of ceramic craftsmanship.
Zhu et al. (Sun,) studied this question.