基于感知哈希的在线发表论文版权保护系统  

Thesis Copyright Protection based on Topic Words Semantic Perception Hash

作　　者：李娇[1] 王健[1] 曹继文[1] 陈彤[1] 

机构地区：[1]北京交通大学计算机与信息技术学院,北京100044

出　　处：《信息网络安全》2013年第11期31-34,共4页Netinfo Security

基　　金：教育部人文社会科学研究规划基金[13YJA870013];2013年北京交通大学大学生创新训练计划项目

摘　　要：针对语义相似词语替换形成的论文抄袭,文章提出了基于主题词语义感知的论文分块感知哈希算法。算法首先使用交叉熵准确提取主题词感知论文内容主题,根据论文结构生成分块感知哈希定位修改位置;然后基于主题词语义相似度计算比较论文相似度,根据注册时间戳、作者ID判断版权归属,保护论文原创作者知识版权;最后采用ECDH和Rijndael算法加密感知哈希防止明文泄露,完成作者、出版发行商双向身份认证。测试表明算法具有鲁棒性、可区分性。In allusion to paper plagiarism replaced with semantic similar words, a thesis segment perceptual hash algorithm based on the topic words semantic perception is proposed in the paper. Firstly, the algorithm extracted topic words from the thesis according to Kullback-Leibler Divergence, sensed the theme of content, built segments perceptual hash on the structure to locate modify. Then compared thesis similarity computed topic words semantic similarity, estimated ownership of copyright used author ID and registered timestamp, protected the author original copyright. Finally it encrypted hash with ECDH and Rijndael algorithm to prevent plaintext leaked and complete mutual authentication between author and publisher. Experimental results show the approach has robustness and discrimination.

关 键 词：感知哈希 语义相似 版权保护 主题词提取 内容安全 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构] TP391.1[自动化与计算机技术—计算机科学与技术]