Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{184916,
        author = {Varna O V and Shreya Sreekumar and Sarang C and Alka Sajeevan P and Subhaga K},
        title = {An Analytical Review of Steganography: Traditional Methods, Emerging Hybrids, and Deep Learning Innovations},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {4},
        pages = {4521-4534},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=184916},
        abstract = {Image steganography is a popular method for hiding information in plain text, but it is still a challenge in digital security due to its high embedding capacity, high fidelity, and robustness against compression. In this paper, we present a novel approach to image hiding that effectively balances three long-standing challenges, namely, embedded capacity and image fidelity. We explore the use of deep convolutional neural networks (DCGANs) to address the dual challenge of maintaining high- capacity hiding and ensuring strong fidelity against modern steganalysis techniques. This paper introduces a hybrid approach that integrates edge detection with Convolutional Generative Adversarial Networks (CNNs), leveraging the strengths of Pixel Value Differencing (PVD) and Least Significant Bit (LSB) substitution with the Histogram of Oriented Gradient (HOG) algorithm. Experimental evaluations across multiple datasets, including DIV2K, COCO, and ImageNet, demonstrate that StegTransX significantly outperforms state-of-the-art methods in single-image and multi-image hiding tasks, achieving higher PSNR, SSIM, and generalization capability while requiring fewer parameters and FLOPs. The results demonstrate that the proposed method achieves competitive PSNR values and better imperceptibility compared to conventional PVD- and LSB-only techniques, demonstrating that edge detection and CNN enhance- ment are a viable direction for concealing data while maintaining visual fidelity while maintaining invisibility. The study explains the mathematical foundation of RSA, particularly its reliance on large prime factorization and modular exponentiation, and demonstrates its effectiveness in ensuring confidentiality and integrity of digital communication.},
        keywords = {Steganography, Information Hiding, Least Signifi- can’t Bit, Transform-Based Methods, Deep Learning, Generative Adversarial Networks, Robustness and Imperceptibility.},
        month = {October},
        }