Dual-stage encryption framework for secure 3D information hiding using single-pixel imaging and diffusion models.

Area of Science:

• Computer Vision
• Information Security
• Digital Imaging

Background:

• Three-dimensional (3D) imaging is vital for medical diagnostics, digital heritage, and virtual reality.
• Protecting sensitive spatial information in 3D data is crucial.
• Existing 3D information hiding methods face challenges in capacity, imperceptibility, and robustness.

Purpose of the Study:

• To develop a novel 3D information hiding method.
• To securely embed depth data into host images without visual distortion.
• To enhance confidentiality, imperceptibility, and recovery accuracy.

Main Methods:

• Integration of single-pixel imaging with diffusion models.
• A dual-stage encryption framework for enhanced security and imperceptibility.
• Adaptive weight allocation strategy prioritizing regions of interest.

Main Results:

• The proposed method significantly outperforms existing approaches.
• Demonstrated superior hiding and recovery performance.
• Showcased strong potential for secure 3D data transmission and storage.

Conclusions:

• The novel 3D information hiding method offers enhanced security and data integrity.
• The integration of advanced imaging and AI techniques addresses key challenges in 3D data protection.
• The method is well-suited for practical applications requiring secure 3D data handling.