Information hiding based on Augmented Reality.

Area of Science:

• Computer Science
• Information Security
• Computer Vision

Background:

• Traditional information hiding methods struggle to balance secret embedding rate and detection accuracy.
• Carrier-based information hiding techniques often exhibit deficiencies in practical applications.

Purpose of the Study:

• To propose a novel information hiding architecture utilizing Augmented Reality (AR) to enhance secret communication.
• To address the limitations of traditional methods by integrating AR, deep learning, and secret-key matching.
• To develop a coverless information hiding scheme with improved embedding capacity and security.

Main Methods:

• Developed an AR-based information hiding architecture combining information hiding, AR, and deep learning.
• Implemented a secret-key matching policy where secret messages are mapped to objects, images, or coordinates.
• Integrated deep learning models, specifically Convolutional Neural Networks (CNNs), for real-time object detection and image recognition.
• Utilized AR to transmit secret messages in various formats, enabling higher potential embedding rates.

Main Results:

• The proposed architecture demonstrated a higher potential secret embedding rate by leveraging AR capabilities.
• The system achieved coverless information hiding, reducing the detectability of hidden data.
• Experimental implementation using CNNs for object detection, image recognition, AR, and secret-key matching showed promising practical results.

Conclusions:

• The novel AR-based information hiding architecture offers a promising solution for secure communication.
• The integration of AR and deep learning significantly enhances the potential for high-capacity and secure information hiding.
• The developed system represents a practical advancement in coverless information hiding techniques.