Data Hide using Deep Neural Network in Encrypted Images

Ms. Nagma Shaikh, Ms. Pooja Waghamare, Ms. Bhavana Potrajwar, Mrs. Rupali Sarde

Data Hiding, GAN model, Deep Neural networks RDH.

In recent day, images are shared on social media platform, for this image security is big problem. In order to conceal the message from the image and vice versa, we would like to employ steganography and coding techniques. We often use a lossless reversible technique for embedding and extracting information within the designed system. By gently altering the pixel values, we can insinuate secret data into the cowl image via a technique known as reversible information concealment. In this paper, in this paper we study alternative approach for combining models such as convolution neural networks and generative adversarial networks to obtain meaningful encrypted images for RDH. In this practical we designed using a four-stage specification t includes the hiding network, the encryption/decryption network, the extractor, and ultimately the recovery network. Through residual learning, the crucial information was incorporated into the image within the concealing network. The main image is encrypted using GAN into a meaningful image called as the embedded image inside the encryption/decryption network. The embedded image is restored to the decrypted image. In this to fully extract the secret message on the receiving end, the original image must be retrieved. The numerous uses, including social control, the medical field (where patient data confidentiality is an example), and the military, where the ability to conceal information is highly valued.

https://doi.org/10.31871/IJNTR.10.2.7