Covert Communication through Robust Fragment Hiding in a Large Number of Images

Area of Science:

• Computer Science
• Information Security
• Digital Watermarking

Background:

• Covert communication in lossy channels faces challenges from image-processing attacks on hidden watermarks.
• Ensuring undistorted secret information extraction after attacks requires robust watermarking techniques.

Purpose of the Study:

• To design a novel fragmented secure communication system for reliable data transmission over lossy channels.
• To develop a strong, robust image watermarking method resistant to various attacks.

Main Methods:

• Fragmenting secret data and redundantly hiding it in numerous multimodal carriers (images) across social platforms.
• Employing a dual-layer watermarking approach: DFT for efficiency and DCT for resistance to translation and JPEG attacks.
• Implementing a fragmentation strategy where each data fragment is embedded in multiple images.

Main Results:

• The proposed watermarking method achieves high embedding/detection efficiency and good invisibility (average 41dB PSNR).
• Embedding and extraction times are rapid, under 0.15s for images up to 2000x2000.
• The system demonstrates high robustness, achieving 100% secret file restoration under various attacks (rotation, scaling, cropping) and significant image loss (up to 30%).

Conclusions:

• The fragmented secure communication system, combined with robust image watermarking, is effective and practical for covert data transmission.
• The method ensures complete secret data restoration even with partial carrier loss or under severe image processing attacks.
• The system offers a significant improvement in robustness and reliability compared to existing schemes.