Image Data Hiding Scheme Using Adversarial Embedding

Data and information are increasingly at risk due to developments in science and technology. Several strategies for hiding information have been proposed to deal with these threats. Hiding information in a picture is the most commonly used method since it is difficult for an outsider to detect hidden information. Data coverage can be improved by using different deep learning and artificial intelligence techniques. This paper proposes a deep learning based framework steganography embedding-extraction architecture for embedding pictures as payloads. To achieve this, we have made some important contributions: (i) the general encoder-coder architecture, which encodes and decodes all sequences of pixels, is proposed;(ii) we are introducing a loss function that ensures the training of embedded extraction networks; (iii) on a range of freely available datasets, we will be doing a thorough analysis of the proposal for architecture; and (iv) to enhance security and quality of image, we are introducing a new adversary network for the detection of stego images. Indicate the high levels of PSNR and SSIM with cutting-edge payload capacities.