Applied Research In Electronic Document Secutity

With the rapid development of electronic information technology and computer networks, information network has covered many fields. The transmission, collection and sharing of information are closely combined with office automation, and thus the people's requirements on document security and authenticity are on the rise. In the course of information transmission and security, current documents, e-mail and the like generated by the office system are facing with the following issues: how to determine the identity of the signer, how to guarantee the integrity of the data, and how to prevent repudiation of the originator. Taking the security issues into account, we think it necessary to develop a security electronic document system.This paper proposes a new method combined with the on-line handwriting signature system to guarantee the integrity of documents and verify a signer's identity. We use MD5 algorithm to examine the integrity of documents and BP neural network to verify the signer's identity. For the human eyes have different sensitivity to green, red, and blue, this paper proposes a hybrid method of LSB algorithm, which is a way to hide information by modifying the minimum significance bit of the carrier image's Gray value.We get the datas of on-line handwriting signature through writing pad, extract the representative handwriting signature characteristics and pack the signer's handwriting signature into the signature image. The message digest generated from the electronic document by MD5 algorithm and the handwriting signature features identifying a signer, are invisibly embedded in the signature image.When received the electronic documents, we extract the handwriting signature features from the signature image which contains hidden information. Then we send them to the trained BP neural network to verify the identity.In the design of the BP network, the improved BP network classifier is used; both the true signatures and the forgery signatures are put into the network to get trained. We determine the expect output value of output neurons layer by comparing of the experiments. Designing the network structrue, and improving the standard BP algorithm are the works of this paper. We also determine the training mode is the single-sample mode.On-line handwriting signature verification uses BP neural network to guarantee reliability of verification. Verification information is hidden in the signature image to improve the security of verification.