Joint Steganography and Encryption Techniques for Security Enhancement Using Deep Learning Models

Joint Steganography and Encryption Techniques for Security Enhancement Using Deep Learning Models

연구 분야: Strategies

학회: IVSP '24: Proceedings of the 2024 6th International Conference on Image, Video and Signal Processing

In today’s rapidly expanding digital landscape, safeguarding sensitive data and privacy has become imperative. Traditional security measures and authentication technologies are facing increasing challenges, particularly steganography. Steganography involves concealing messages within various forms of media, such as images, audio, or video, to avoid detection by third parties. Unfortunately, conventional steganography techniques are vulnerable to advanced decryption methods, emphasizing the urgent need for more robust strategies. The proposed scheme introduces an innovative approach to conceal secrets through adversarial reinforcement and diffusion models. In this model, the secret image transforms noises, and an inverse model is designed to revert the noise to the original secret image. Experimental results demonstrate the feasibility and effectiveness of the proposed scheme.