Implementation and Analysis of Image Steganography using Convolution Neural Networks

Steganography is an art of hiding the data which is used for communication purposes, and it is not visible to the human eye. By hiding information with the help of other information communication takes place. The hiding information is in different forms like text, image, audio, and video. Image Steganography is performed for images, and it takes two images as input, and they are combined by using an embedding algorithm and the data is decrypted to retrieve the original image. Image Steganography is classified into three types General Adversarial Network methods, Convolutional Neural Network (CNN) based, and traditional methods. The traditional method is based on Least Significant bit method it gives results in less time. But it is unprotected to do steganalysis, and it is not secure at all. For this purpose, we move to the Convolutional Neural Networks. In this CNN based deep learning Neural Networks are used. The comparison of input and output images is done by considering parameters like Histogram, MSE, SNR, and PSNR.