Evaluating the Robustness of Image Steganography Techniques against Diverse Degradation Techniques

Abstract

Within the field of steganography, image steganography is a fascinating technique that hides confidential data behind a cover image. However, the quality and security of hidden content are threatened by potential degradation introduced by the transmis sion of stego images, which can take the form of noise, blurring, or sharpening. In this research endeavor, we present an innovative approach by integrating image degrada tion models into an existing state-of-the-art invertible image-in-image steganography model known as HiNet. Our proposed methodology involves applying a degradation model to the stego image post-secret image embedding, followed by the usual secret image extraction. By introducing these additional layers to the steganographic pro cess, we aim to enhance the robustness of our model against various degradation sce narios, such as Gaussian noise, blurring, and sharpening. By using knowledge from cutting-edge architectures such as HiNet, we aim to improve the overall security and quality of hidden pictures. We performed experiments on HiNet and achieved im proved results in handling degradations like noise, blurring, and sharpening.