Publication

Abstract : Steganography is a methodology that essentially aims in hiding a secret data in a fitting carrier, essentially in such a way that the hidden data does not attract any attention toward it. The postulation in steganography is if the secret feature is perceptible, then so is the point of attack. Also, clean images when subjected to prolonged transmission, improper image acquisition or conditioned to multiple feature changes lead to image tarnishing due to unwanted noisy pixels. This proposes to be a major threat in steganography as the presence of noise itself could be a reason for the secret image to be revealed. In this work, a bi-objective end-to-end stacked denoising stenography model is implemented. The results are estimated by evaluating the average peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) of the trained images using COCO dataset. The stability of the proposed model is estimated using the average loss obtained by training the model for multiple iterations.