Multi-Modal Steganography for Secure Data Hiding across Diverse Digital Carriers

Abstract

Secure communication today must satisfy two fundamental objectives simultaneously: the content of a message must remain protected, and the communication itself should not appear suspicious. Conventional cryptographic methods such as AES, RSA, and DES are highly effective at protecting data confidentiality, but they leave encrypted traffic visible and therefore identifiable. In many environments, this visibility alone is sufficient to attract monitoring, traffic analysis, blocking, or even forced disclosure. Steganography addresses this limitation by concealing the existence of the message itself within ordinary digital carriers. This report presents a multi-modal steganography framework for secure data hiding across images, audio files, PDFs, and videos. The system follows a layered design. First, the payload is encrypted using AES-256-GCM, providing confidentiality, integrity, and authentication. Second, the encrypted bitstream is embedded using a hybrid architectural approach: spatial and temporal randomized least significant bit (LSB) substitution, driven by a pseudo-random number generator seeded from secret material, is applied to continuous media streams (images, audio, and video), whereas structural object stream parsing and metadata injection are employed for PDF document carriers. Third, the receiver extracts the embedded payload, verifies the authentication tag, and decrypts the message only when the correct key is provided. This design reduces the effectiveness of basic steganalysis because the hidden data is distributed non-sequentially, and the recovered payload remains unintelligible without the encryption key. The paper further discusses carrier suitability, implementation choices, capacity analysis, and evaluation methodology. Performance assessment is based on PSNR, SSIM, MSE, SNR, payload capacity, and detectability considerations.

Introduction

This text describes a multi-modal secure data hiding framework that combines cryptography and steganography to protect confidential digital communication. Modern digital communication makes sharing information fast and convenient, but it also increases the risk of interception through networks, cloud systems, and shared platforms. Encryption protects the content of a message by converting it into unreadable ciphertext, but it does not hide the fact that communication has occurred. Steganography addresses this limitation by concealing the existence of the message inside ordinary-looking files such as images, audio, videos, or documents.

The proposed framework uses a layered security approach:

1. AES-256-GCM encryption is applied first to protect message confidentiality and integrity.
2. The encrypted payload is then hidden inside a suitable carrier file using different embedding techniques depending on the media type.

The framework aims to overcome limitations of traditional systems, such as:

• Visible encrypted traffic that may attract attention.
• Single-carrier limitations.
• Detectable patterns caused by simple sequential embedding methods.
• Poor payload management and carrier distortion.

Cryptographic Foundation

The system follows the principle of encryption before steganography. If hidden data is extracted, attackers still obtain encrypted information rather than the original message. The framework uses AES-256-GCM, which provides:

• Strong encryption.
• Authentication through an integrity tag.
• Protection against message tampering.

For password-based use, the system applies key derivation techniques such as PBKDF2 with salts and multiple iterations to prevent brute-force attacks.

Steganographic Methodology

The main hiding technique is Least Significant Bit (LSB) embedding, where the smallest bits of digital media samples are modified to store secret information. Since these changes are very small, they are usually not noticeable.

However, simple sequential LSB insertion can create detectable patterns. To improve security, the framework uses pseudo-random number generator (PRNG)-based embedding, which randomly selects storage locations using secret key-derived information. This makes extraction difficult for unauthorized users.

Different carriers are handled differently:

• Images (PNG/BMP): Use randomized LSB embedding due to tolerance for small pixel changes.
• Audio (WAV): Uses careful sample modification while maintaining sound quality.
• PDF documents: Stores hidden data in internal structures such as metadata and object streams rather than visible content.
• Videos (MP4/AVI): Uses spatial and temporal redundancy for larger payload capacity.

System Architecture

The framework consists of two main pipelines:

Sender Pipeline:

• User selects a secret message and carrier file.
• System checks whether the carrier has enough capacity.
• Payload is encrypted using AES-256-GCM.
• Ciphertext is embedded into the carrier using randomized techniques.
• A normal-looking stego file is produced.

Receiver Pipeline:

• Receiver extracts hidden bits using the correct secret key and PRNG sequence.
• The ciphertext is reconstructed.
• Authentication is verified.
• The original message is decrypted only if integrity checks succeed.

Conclusion

This report presents a structured multi-modal steganography framework for secure data hiding across diverse digital carriers. The design combines authenticated encryption, randomized embedding, carrier-aware capacity checking, and flexible support for images, audio, PDFs, and videos. The resulting system does more than conceal content; it conceals communication itself while still preserving recovery integrity. The major strength of the framework is its layered approach. AES-256-GCM protects the payload even if it is found. Randomized LSB insertion reduces predictable patterns. Carrier selection prevents misuse. Together these features create a practical covert communication model that is suitable for real-world secure data transfer scenarios where simple encryption is not enough. The contribution of the work is therefore both technical and practical. Technically, it combines encryption and steganography in a clean pipeline. Practically, it allows the same security logic to be reused across multiple file formats, which increases deployment flexibility.

References

[1] I. Haverkamp and D. K. Sarmah, “Evaluating the merits and constraints of cryptography–steganography fusion: A systematic analysis,” Int. J. Inf. Secur., vol. 23, pp. 2607–2635, May 2024. [2] R. Klemm and B. Chen, “Hiding sensitive information using PDF steganography,” arXiv preprint arXiv:2405.00865, May 2024. [3] S. Rahman, J. Uddin, H. Hussain, S. Shah, A. Salam, F. Amin, I. de la Torre Díez, D. L. Ramírez Vargas, and J. C. Martínez Espinosa, “A novel and efficient digital image steganography technique using least significant bit substitution,” Sci. Rep., vol. 15, Art. no. 107, Jan. 2025. [4] T. Radivilova, I. Dobrynin, A. Snihurov, S. Stanhei, and S. Bulba, “Image steganography method using LSB and AES encryption algorithm,” in Proc. CQPC 2025, CEUR Workshop Proceedings, Aug. 2025. [5] D. T. Firdaus, N. J. D. C. Ntivuguruzwa, T. Ahmad, D. Mukanyiligira, and L. Sibomana, “Steganographic model to conceal the secret data in audio files utilizing a fourfold paradigm: Interpolation, multi-layering, optimized sample space, and smoothing,” J. Saf. Sci. Resil., vol. 6, no. 2, pp. 138–149, Jun. 2025. [6] D. Kumar, V. K. Sudha, N. Manikandan, and K. Ramaswamy, “Efficient three-layer secured adaptive video steganography method using chaotic dynamic systems,” Sci. Rep., vol. 14, Art. no. 18301, Aug. 2024. [7] M. ud Din, I. A. Shoukat, E. Arif, M. Amjad, M. Mohsin, M. Mumtaz, and M. A. Rauf, “Randomized frame selection-based video steganography method for secure embedding of secret data,” J. Comput. Biomed. Inform., vol. 8, no. 2, 2025. [8] O. Omego and M. Bosy, “Multichannel steganography: A provably secure hybrid steganographic model for secure communication,” arXiv preprint arXiv:2501.04511, Jul. 2025. [9] H. El-Taj, “A secure fusion: Elliptic curve encryption integrated with LSB steganography for hidden communication,” Int. J. Comput. Exp. Sci. Eng., vol. 10, no. 3, pp. 434–460, 2024, doi: 10.22399/ijcesen.382. [10] M. A. Nasr, W. El-Shafai, E.-S. M. El-Rabaie, A. S. El-Fishawy, H. M. El-Hoseny, F. E. Abd El-Samie, and N. Abdel-Salam, “A robust audio steganography technique based on image encryption using different chaotic maps,” Sci. Rep., vol. 14, Art. no. 22054, 2024, doi: 10.1038/s41598-024-70940-3. [11] S. Z. Banoori, W. Khan, S. Rahman, F. Masood, A. Salam, F. Amin, I. de la Torre, M. G. Villar, H. Garay, and G. S. Choi, “An improved hybrid image steganography method using AES algorithm,” Sci. Rep., vol. 15, Art. no. 44515, 2025, doi: 10.1038/s41598-025-28140-0. [12] J. Singh, A. Kumar Singh, and S. S. Chauhan, “Enhanced edge-based steganography using image segmentation and random LSB substitution for secure data hiding,” Int. J. Smart Sens. Intell. Syst., vol. 18, no. 1, 2025, doi: 10.2478/ijssis-2025-0049. [13] M. Helmy and H. Torkey, “Secured audio framework based on chaotic-steganography algorithm for Internet of Things systems,” Computers, vol. 14, no. 6, Art. no. 207, 2025, doi: 10.3390/computers14060207. [14] A. Aravind Krishnan, Y. Ramesh, U. Urs, and M. Arakeri, “Audio-in-image steganography using analysis and resynthesis sound spectrograph,” IEEE Access, pp. 75180–75189, 2025, doi: 10.1109/ACCESS.2025.3563781. [15] S. N. Al-Rekaby, M. A. Khodher, and L. K. Adday, “Video steganography using a 9-D chaotic system and an encryption algorithm for data hiding,” Int. J. Saf. Sec. Eng., pp. 1007–1018, 2025, doi: 10.18280/ijsse.150514.

Copyright

Copyright © 2026 Prof. P. V. Sridevi, Allaka Nitin, Palla Yogith Yadav, Gandi Pavan Naga Venkatesh, Pillala Sridevi. This 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.