Single secret image sharing scheme using neural cryptography

Abstract

The goal of the secret sharing is to share a secret information without any leakage to others. In this paper, we proposed a secure mechanism of sharing secret shares of an image between two parties generated via Shamir’s scheme with the help of neural cryptography. Neural cryptography is a new source for public key cryptography schemes which are not based on number theory, and have less computation time and memory complexities. Neural cryptography can be used to generate a common secret key between two parties. Keeping all these in mind, our main focus is to share secret information over a public channel with less computation power. In the case of neural cryptography, both the parties receive an identical input vector, generate an output bit and are trained based on the output bit. The dynamics of the two networks and their weight vectors is found to exhibit a synchronization state with identical weights. These identical weights are acts as a common key between two parties. The proposed scheme is secure and does not reveal secret information.