Watermarking Applied Research In Electronic Document Identification

With the rapid development of the computer and communication technology, the information network proliferates all over the world. The sharing, transmission and collection of information are closely related to office automation system, office automation software is widely used. But people are disturbed by the security problems when they enjoy the efficient and convenient office automation. At present, it is the crux of the problem to safeguard effectively the authenticity, security and identifiability of information in transmitting office automation information. Based on the above issues, it is extremely necessary to develop an effective identification system.Combined with the electronic document system, an identification algorithm based on on-line handwriting signature features was proposed in this paper. The algorithm implements identification by invisibly embedding on-line handwriting signature features of the originator into the the electronic document host image.We apply the writing pad to obtain the on-line handwriting signature information. Primarily, select 15 most representative signature features and normalize them, then embed the signature features invisibly into the host image according to digital watermark algorithm. Finally send the electronic document with the watermarked image to the receiver.After reviving the electronic document, the receiver extract the signature features form watermarked image to verify the sender's identity by BP neural network, If the certification results is ture,the identity of the sender is genuine, or it is forged.In this paper, the traditional DWT (Discrete Wavelet Transform) watermarking algorithm is improved. A watermarking algorithm based on image segmentation and HVS (Human Visual System Model) is proposed. In addition to embed signature features in different sub-blocks to ensure the independence of features, the algorithm also promotes the robustness of watermark by embedding repeatly the important information of signature feature watermark.The experimental section of this paper analyzes the effect of using the on-line handwriting signature feature watermark, and the parameter declaration is given when the balance between invisibility of watermark and robustness of watermark is achieved.The experimental results show that the watermarking algorithm based on on-line handwriting signature features has a more advantage under the electronic documents compared with other watermarking algorithms.