A Novel Approach to Steganography and its Countermeasures Against Deep Learning and MITM Attacks

Steganography, the old art of hiding a message into the digital format has took a lot of changes due to modern technology. The recent advancement of AI along with Deep Learning took a drastic turn in the act of steganalysis. Steganalysis is the reverse of steganography, where Steganography is the science of hiding the data into image Steganalysis is the art of unrevealing the hidden data from the stego image. This advancement is a boon for steganalysis, but it completely shatters the steganography encryption. This paper introduces a new modern method to securely the concealed image. It combines Steganography with various advanced Techniques like image scrambling, image splitting and fragmentation. This multi-layered approach is intended to thwart contemporary threats, such as attacks utilizing deep learning-based detection and Man-in-the-Middle attacks (MITM). By breaking the image up into pieces and scrambling the pixels, the technique operates. Several encrypted channels are then used to transmit these pieces, making it very difficult for attackers to intercept and recreate the original picture. This results when the attacker even if he gets the fragments of message, he cannot recreate the complete message as the fragments are the key itself which makes this approach unique and robust as each fragment which is send is a key to the complete message. Particularly in high-stakes communication situations, this novel approach seeks to ensure secure and effective image transmission by offering strong protection for sensitive visual data compared to old steganography techniques this method ensures protection against Artificial Intelligence.