In carrying out reversible image steganography, the Generative Adversarial Networks (GANs-based) models have proven to be the most suitable deep learning models for image steganography. Image steganography is a steganography system that hides secret data in an image cover medium without arousing suspicion, and it is defined by the ability to reconstruct the cover medium with no visible distortion after the steganography system has been decoded by extracting the hidden data. In this study, we try achieve the encoding phase in image steganography, where two GAN-base models (CycleGAN and DCGAN) were proposed. Empirical analysis was done to determine a better model for the encoding of image steganography. The Peak Signal-to-Noise Ratio (PSNR), the Structural Similarity Index Metric (SSIM), and bit per pixel (bpp) were used as the metrics for the analysis. The outcome of DCGAN yielded (SSIM=0.48; PSNR=19.86; bpp=24.79) and the outcome of using CycleGAN yielded (SSIM=0.97; PSNR=41.45; bpp=24.97). These values concluded that the CycleGAN was preferable over the DCGAN. Hence, the CycleGAN was adopted as the encoding model.

Reversible image steganography is a steganography system that hides secret data in an image cover medium without arousing suspicion, and it is defined by the ability to reconstruct the cover medium with no visible distortion after the steganography system has been decoded by extracting the hidden data. The most common issue with image steganography systems is a lack of payload capacity, which is caused by the inappropriate selection of a cover medium. In this paper, a solution is proposed by employing Histogram of Pattern Set to prepare the image dataset for use in training the Support Vector Machine (SVM) and Deep Neural Network (DNN) algorithms. The precision values for SVM and DNN were revealed to be 83.39 and 98.11, respectively; the recall values for SVM and DNN were discovered to be 81.85 and 96.29, respectively; and the F-score values for SVM and DNN were discovered to be 82.61 and 97.19, respectively. Because the DNN selection model outperforms the SVM selection model, it was chosen as the cover image selection model for the implementation of reversible image steganography.