Security enhancement in image steganography using neural networks and visual cryptography

Abstract

Data Security has been an essential area of concern as the communication medium is widely used over internet. Image steganography indicates sending the secret data in the cover image while hiding the very existence of it self. Before embedding the secret data in the cover image, it is encrypted in order to obtain better level of security. Steganalysis is concerned with field of defeating steganographic techniques. Hence continuous research has been under taken in order to enhance data security. Cryptography involves a process called encryption and is not concerned with hiding the secret data in the cover image. Visual cryptography is a renowned technique to protect data which is image based. During encryption, the image is split into ‘n’ number of shares. During decryption, these shares are stacked together to get original image. Hence specific decryption algorithm is not necessary and just human visual system would suffice. In order to enhance security mechanism, visual cryptography and image steganography are used together. Neural networks are concerned with identifying the best locations in host image in order to embed the secret data thus improving the image quality.