A Novel Steganography Approach Using S-CycleGAN With an Improvement of Loss Function

Abstract

Technologies related to the internet of things (IoT) have been extensively employed in business operations, military research, networking technologies, etc. With the proliferation of the IoT, data theft and leakage have gradually increased because the data communication system is public. Various novel steganography algorithms have been proposed for data hiding. But in this process, the quality of the hidden image decreases badly. We want to give an accurate color to the image instead of severe color changes. To change it, we have to perform a colorization process by mapping the original image to the stego image. It gains more characteristics according to the actual embodiment. Here, the objective is to improve the appearance by calculating the loss function. The stego image can withstand steganalysis by using the authors' proposed scheme by maintaining its integrity while to some extent improving the quality of the image. The experimental results confirmed the proposed method outperforms the existing methods (SGAN and CycleGAN) for information hiding with enhanced image transmission quality.