Multi-viewpoint tampering detection for integral imaging.

Area of Science:

• Optics and Photonics
• Computer Vision
• Digital Image Processing

Background:

• Current integral imaging systems lack security features, making them susceptible to image falsification.
• High computational demands hinder the practical application of existing 3D integral imaging techniques.

Purpose of the Study:

• To develop a secure 3D integral imaging scheme with enhanced tampering detection and self-recovery capabilities.
• To maintain clear light field display while addressing security vulnerabilities and computational costs.

Main Methods:

• A novel pixel mapping and deep learning approach co-extracts depth and angular data for pixel-wise analysis.
• Region of interest regulation of the 3D light field enables initial verification.
• Multi-viewpoint recovery information is embedded for elemental image array reconstruction.

Main Results:

• The proposed scheme effectively identifies and recovers tampered regions within the 3D light field.
• Experimental results show significant parallax effects and successful double recovery of altered image data.
• The method demonstrates effective tampering detection and multi-perspective recovery.

Conclusions:

• The developed 3D integral imaging scheme provides a secure and efficient solution for light field display.
• This approach significantly improves the robustness of integral imaging against falsification.
• The technique offers a promising direction for secure 3D imaging applications.