A Robust Linked List Steganography Without Embedding for Digital Images

Image steganography technology has received widespread attention in recent years, aiming to hide data in digital images for dissemination. It has been applied in multiple fields such as military, medical, and social media. Steganography without embedding (SWE) technology presents an effective means to avoid image distortion while enhancing privacy. This article proposes a novel and robust SWE method, which effectively improves the anti-attack ability of linked list embedding methods. Specifically, we obtain stable features by performing multiple median downsampling and gradient calculation operations on the cover image and then reconstructing them into a multi-headed unidirectional linked list structure. By using the proposed optimal linked list search algorithm, it is possible to find the corresponding perfect linked list based on provided message fragments. The recipient only needs to use the location information of the perfect linked list to restore the secret message from the carrier image. Remarkably, the proposed method imposes no requirements or modifications on the carrier images, theoretically making it undetectable by steganalysis detection. In addition, the introduction of multiple sampling and gradient computing enables this scheme to maintain extremely high restoration accuracy against various image attacks and interferences. Finally, an analysis and comparison of the steganography capacity of the proposed scheme were conducted, along with an examination of its efficacy.