Jul 2026· International Journal for Research in Applied Science and Engineering Technology· 0 citations
TL;DR
VisionStego is proposed, a duallayer security architecture that pairs symmetric-key encryption with an artificial-intelligence-guided steganographic embedding stage, so that cloud-hosted data is protected in both substance and appearance.
Abstract
Modern cloud platforms now carry a substantial share of the world's sensitive digital traffic, and defending that traffic
calls for more than hiding its meaning - an intercepted ciphertext is still visibly a target worth attacking. Encryption alone
protects content; it does nothing to disguise the fact that a secret is being sent at all. This paper proposes VisionStego, a duallayer security architecture that pairs symmetric-key encryption with an artificial-intelligence-guided steganographic embedding
stage, so that cloud-hosted data is protected in both substance and appearance. The secret payload is first reduced in size
through wavelet-based compression, then encrypted with a shared symmetric key, and finally concealed inside a cover image at
positions chosen by a trained Convolutional Neural Network (CNN) rather than by a fixed or pseudo-random rule. The network
scores each candidate pixel for embedding suitability using edge strength, local variance, and texture complexity, concentrating
modification in regions where it is least visible to the eye and least anomalous to statistical steganalysis. Tested on the Lena,
Baboon, and Peppers benchmark images, the resulting stego images reach a Peak Signal-to-Noise Ratio (PSNR) of up to 46.2
dB and a Structural Similarity Index (SSIM) of 0.985, exceeding conventional LSB, adaptive LSB, and wavelet-based baselines
on every image and every metric tested. These findings indicate that VisionStego offers a practical route to the combined
concealment and confidentiality that neither cryptography nor steganography can deliver in isolation.
Steganography is the practice of hiding secret information within digital media such as images, audio, or video. It
ensures confidential communication by concealing the existence of data itself, unlike encryption. Modern research focuses on
improving invisibility, security, and resistance to detection using AI and deep learning. This project aims to design a secure and
intelligent image steganography system using a hybrid Transformer model. It focuses on increasing data hiding capacity while
maintaining image quality and reducing detectability. The system will be capable of resisting steganalysis attacks, compression,
and noise distortions in real-world usage. Recent studies show that deep-learning steganography models still suffer from low
robustness and poor scalability. Most approaches fail under compression or noise, and their hidden data can be detected by
advanced AI models. High computational cost, limited payload capacity, and dataset dependency further affect their reliability.
This project introduces a hybrid Transformer integrated with Discrete Cosine Transform (DCT) for frequency embedding. By
combining spatial and frequency domains, it ensures better concealment and robustness. Adversarial training with a steganalysis
discriminator will enhance security against modern detection models. The system will achieve higher PSNR and SSIM scores,
proving superior imperceptibility and accuracy.
Implementation will use Python, PyTorch, and OpenCV for model training and image processing. Datasets like COCO,
BOSSBase, and ImageNet will be used for evaluation. Performance metrics such as PSNR, SSIM, MSE, and BER will measure
quality and accuracy. Development and testing will be carried out in Jupyter Notebook or Google Colab environments.
Aakash Bonagiri, N. N. Kumar· International Journal for Re...· 0 citations
A new framework that combines hybrid encryption with saliency-based adaptive embedding to select the most effective regions for data concealment in cover images, based on the BossBase dataset is proposed.
Abdullah S. al-Malaise Alghamdi, Rana Alrawashdeh· Journal of Cyber Security an...· 0 citations
Secure transmission of digital video must Coordinate two competing goals: protecting the content against cryptographic attacks while preserving format compliance and low complexity for real-time multimedia delivery. This paper presents an integrated joint video coding-and-encryption framework in which a convolutional neural network (CNN) cipher is embedded into a modified MPEG-2 codec. The cipher operates as a keyed value transformation whose weights and biases are generated from a chaotic logistic map, with the control parameter μ and the initial value x₀ acting as the secret key encryption is applied in the coded domain to the quantized transform coefficients and motion vectors before entropy coding. Two operating modes are supported: a full-protection mode that encrypts coefficients and motion vectors, and a perceptual mode that encrypts only the motion-vector stream together with a reduced quality factor, yielding a controllable low-quality preview suitable for try-before-buy services. Building on a focused review of prior work on known-plaintext attacks (KPA) and perceptual encryption, the scheme is assessed against the KPA threat through an XOR-based analysis, which shows that the effective mask recovered from one plaintext–ciphertext pair cannot decrypt a different encrypted scene and instead produces a noise-like result. Experimental results demonstrate that the proposed framework resists the known-plaintext attack while providing controllable perceptual protection, confirming its suitability for secure and efficient video transmission.
T. Fadil, Nagham Hamid, Mohammed Alshaikha Ali· Al-Noor Journal of Engineeri...· 0 citations
Reversible data hiding in encrypted images (RDHEI) often creates embedding room by preserving or preprocessing image redundancy, thereby coupling the hiding layer to a specialized encryption model. This paper presents a dual-image method that works directly on AES-CTR ciphertext. Its design rests on four elements: a keyed Sudoku coordinate code for carrying two base-8 digits per accepted pair; exact reconstruction of each ciphertext pair from inter-block displacement; AES-GCM framing for payload confidentiality and integrity; and an adaptive movement threshold that selects the lowest-distortion mappings able to accommodate the complete frame. The data hider requires neither plaintext nor the image-encryption key, and no location map is transmitted. On sixteen 512×512 grayscale images, the method achieves a mean maximum net rate of 1.4187 bits per transmitted pixel. At ERt=1.0, the two marked outputs reach mean PSNR values of 47.65 dB, and all 160 image-rate trials yield zero payload error and pixel-exact recovery. An independent implementation of the embedding and recovery layer of Venkatesh et al. is also examined; published and reproduced results are reported separately where the original specification leaves ambiguity. The experiments show that the proposed architecture provides predictable payload, low carrier-domain distortion, authenticated framing, and exact recovery while retaining standard image encryption.
Cao Thi Luyen· Journal of Science and Techn...· 0 citations
Coverless image steganography hides secret information without directly changing the cover image. Because the original image is not modified, it is more difficult for image steganalysis tools to detect the hidden information. In this research, Wasserstein Generative Adversarial Networks (WGAN) and Local Binary Pattern (LBP) features are used to securely hide and transmit sensitive information using a single cover image. The proposed method provides higher hiding capacity and better security than traditional methods that require multiple cover images.First, the cover image is divided into overlapping blocks. The LBP feature is calculated for each block and converted into a hash code. The secret message is then matched with image blocks having the same hash codes. A lookup table is used to make the embedding process faster.The generated stego information is given to a GAN model to create a meaningful but unrelated image. This image can be sent to the receiver instead of directly sending the stego image. At the receiver side, the WGAN model is used to reconstruct the required stego information. The use of overlapping blocks helps generate many unique hash codes. Without overlapping blocks, a single image may not provide enough hash codes to hide the complete secret message.
Vijaysinh Jadeja, Khyati Rami, Swati Patel et al.· International journal of com...· 0 citations
Protecting confidential information requires not only preventing unauthorized access to encrypted data but also concealing the very existence of the protected information. In this work, we propose a computational dual-layer optical security framework that integrates optical encryption and steganographic camouflage into a unified strategy. The proposed method is based on a 4F optical architecture and employs two two-dimensional private keys: a phase-only key represented through Circular Harmonic Components (CHC) and a periodic amplitude mask acting as a second secret key. Their combined action generates visually diverse steganograms from encrypted RGB images while preserving the correct recovery of the original information by authorized users. Unlike conventional optical encryption methods that produce easily recognizable cryptograms, the proposed approach disguises the encrypted information within camouflage patterns, providing an additional layer of protection before any decryption process is attempted. Numerical simulations demonstrate successful encryption, camouflage, and image recovery while showing that different steganographic appearances can be generated by modifying the private key and the periodic-mask parameters. Furthermore, a prospective optical implementation based on a Mach--Zehnder interferometer and digital holographic recording is presented, providing a feasible path toward future experimental realization. The proposed methodology is introduced as a proof of concept of the dual-layer optical security framework; a comprehensive cryptanalytic evaluation is beyond the scope of this first study and is left for future work.
Jorge-Enrique Rueda-P, C. Pinzón· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
