Image steganography techniques for resisting statistical steganalysis attacks: A systematic literature review

Area of Science:

• Computer Science
• Information Security
• Cybersecurity

Background:

• Information hiding in images, or image steganography, is increasingly relevant.
• Statistical steganalysis techniques are emerging to detect hidden messages, making images vulnerable to cyber-attacks.
• Existing research on robust image steganography techniques is limited.

Purpose of the Study:

• To provide a comprehensive systematic literature review of image steganography schemes resistant to statistical steganalysis attacks.
• To update researchers and practitioners on current trends in information security assurance mechanisms.
• To address the gap in literature concerning advanced image steganography.

Main Methods:

• Systematic literature review of 125 articles from major digital libraries (ACM, IEEE, Science Direct, Wiley).
• Synthesis and analysis using PRISMA methodology with quantitative and qualitative approaches.
• Discussion of techniques' robustness against statistical steganalysis (e.g., RS, Chi-Square).

Main Results:

• Generative Adversarial Networks (GANs) are dominating current image steganography research.
• Artificial intelligence (AI) algorithms like Machine Learning, Deep Learning, and Convolutional Neural Networks enhance security.
• Traditional methods like Least Significant Bit (LSB) algorithms are receiving less attention.

Conclusions:

• AI and GANs represent the forefront of secure image steganography techniques.
• Future research should explore emerging technologies like blockchain and biometrics for enhanced steganography.
• The study highlights the evolving landscape of information hiding and its security implications.