Enhancing Image Security Using DNA and Deep Steganography

This project is aimed at providing a secure way of sending confidential images between users. Application is built to provide protection of the sensitive image being transmitted. Modern advancements have made it incredibly convenient to both send and store digital images online. Images often contain sensitive personal information, such as photographs, identifica- tion documents, medical records etc. Unauthorized access to these images can lead to identity theft, privacy invasion, and personal harm. Image security is crucial for several reasons, spanning personal privacy, corporate confidentiality, intellectual property protection and national secu rity. In order to solve the above described problems, this project provide a novel encryption method for color pictures that utilizes convolutional autoen- coder, DNA and steganography. Convolutional autoencoders can compress image data into a lower dimensional latent space before encryption. Image encryption scheme using DNA along with multiple chaotic sequences and substitution boxes provide security of the image. Hiding images within other images involves a technique known as steganography. In this process a secret image is embedded within a cover image making it imperceptible to the human eye. The security of the proposed method has been evaluated using several parameters, such as histogram of the cipher image, entropy, NPCR, UACI etc. The result shows that it a compelling approach to secure images against various attacks. Keywords− Convolutional neural network, autoencoder, chaos, DNA coding, deep learning, colour image encryption, steganography