Data encryption Using Backpropagation Neural Network

Abstract

The aim of this research is to build a ciphering technique by using artificial neural network to protect data against unauthorized access to the data being transferred.The encryption data includes three stages: first Stage :- Training a network by using backpropagation to obtain weights. Second Stage:- Encryption data by using the weights obtained from first stage and consider the weights of first layer as a public key. third Stage:- Decryption data by using the weights obtained from the first stage and consider the weights of second layer as a private key. The three stages are attained 100% success for data encryption process and data getting back process.This technique is similar to coding asymmetric, and have the ability of coding a group of data such as:- text, characters, numbers and waves. This work is executed by computer type P4 with whole equipments and Matlab language version 7.