Novel Techniques of Data Hiding in HTML Documents for Copyright Protection, Tampering Detection, and Covert Communication

• Main Authors: Pao-Hsun Lai, 賴保勳
• Other Authors: Da-Chun Wu
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/26333362318610672051

碩士 === 國立高雄第一科技大學 === 資訊管理所 === 92 === The rapid advance of technology has made the Internet thrive. Web page is one of the main sources on the Internet people use for both searching and spreading information as well as the most popular service among the others offered. A web page is an HTML document. Via Internet's immediacy, we could use it to send information. A HTML document is easy to duplicate, so it might be misused and even tampered causing the depreciation of the web's credibility. So, this dissertation will introduce some methods used especially in HTML documents to do information hiding in order to achieve covert communication, copyright protection, and tamper detection. The information hiding technique we offer here won't affect how the HTML will look in a web-browser after embedded information. The hiding technique uses Tag to hide information. On the behalf of covert communication, we use Attribute With Default Value and Permutation of Attributes to hide secret information within Tag. After hiding, documents could still be written according to W3C hence quite un-suspect, hidden, and safe. On the behalf of copyright protection, we use five different information hiding techniques to embed the copyright signals. Each technique could embed many copies of copyright and in the mean time the signals in each technique won't affect each other. During extraction process, an additional voting scheme could increase copyright signal's accuracy. And if some of the signals were wrecked, the correct copyright would still be defined from the remaining signals. At last, two tamper detection schemes are proposed which are temper proofing with HTML source codes and webpage displays by using mechanisms of block authentication and dummy signals respectively. First we divide HTML into un-equal blocks. Each block calculates its authentication signal. During the embedding process, we find each dummy signal that is corresponding to its specific authentication signal and embed it into the block it belongs. Experiment proves this way can efficiently detect where has been tampered. Moreover, this dissertation suggests implementation of Active Tamper Detection Agent. When a web page is being tampered in the server, the tamper detection agent will stop the web page service immediately to avoid wrong information spreading via the net.