A Novel Robust Image Encryption Scheme Based on Hybrid Autoencoders and Generative Adversarial Network (GAN)

Hybrid Transformer-Based Steganography Using Frequency Embedding and Adversarial Robustness

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.

Quantum-resilient image steganography with MitM resistance via symmetry-based compression and TransformerGAN

VisionStego: An AI-Guided Dual-Layer Cryptographic Steganography Framework for Secure Cloud Data Transmission

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.

Saliency-Aware Steganography with a Hybrid AES and RSA with LSB Embedding

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.

RIS-MoE: robust and secure image steganography via latent-space optimization with mixture-of-experts denoising

Diffusion-based generative image steganography enables covert communication by synthesizing stego images without relying on cover images. However, existing latent-space methods still struggle to balance robustness, steganographic security, and visual fidelity, especially under practical channel distortions such as compression, blur, resizing, and noise. To address these challenges, we propose RIS-MoE, a robust and secure latent-space image steganography framework that integrates distortion-tolerant message representation with receiver-side adaptive latent restoration. At the sender side, a learnable orthogonal transformation converts the secret message into a distributed representation, which is embedded into the diffusion latent through a residual-guided Hide Network. At the receiver side, a plug-and-play Mixture-of-Experts (MoE) denoising module estimates the distortion composition and adaptively fuses specialized restoration experts before message extraction. Extensive experiments show that RIS-MoE achieves strong robustness under single, mixed, and real-world distortions. It maintains extraction accuracy above 90% under all evaluated simulated combined distortions and achieves 94.62% and 95.29% extraction accuracy after real-world Weibo and Instagram transmission, respectively. RIS-MoE also achieves competitive empirical resistance against spatial-domain, latent-domain, and diffusion-aware steganalyzers, while maintaining favorable visual quality with an FID of 7.35 and an LPIPS of 0.21 on Flickr8K. In addition, the proposed MoE module consistently improves representative latent-space steganography pipelines as a plug-and-play restoration component, demonstrating its transferability. The source code is publicly available at: https://github.com/angle-cell/RIS_MOE.

Video Steganography Based on Invertible Neural Networks With DenseNet and U-Net Features

Comparative results with existing state-of-the-art methods show that this end-to-end embedding approach delivers promising performance in terms of human visual accuracy, with a PSNR score of 37.054 and an SSIM score of 0.9754.