基于加性噪声的强鲁棒数据库盲水印方法  

A robust blind watermarking method for databases based on additive noise

作　　者：李程 张国栋[1] Li Cheng;Zhang Guodong(Unit 96713 of the People′s Liberation Army,Shangrao 334000,China)

机构地区：[1]中国人民解放军96713部队,江西上饶334000

出　　处：《网络安全与数据治理》2025年第2期32-38,43,共8页CYBER SECURITY AND DATA GOVERNANCE

摘　　要：针对大数据共享和发布过程中面临的侵权盗版与隐私泄露问题,提出了一种基于加性噪声的强鲁棒数据库盲水印方法。该方法结合了数据库水印与数据脱敏技术,利用加性拉普拉斯噪声已有失真,通过比值奇偶性与噪声修改机制嵌入水印信息,显著减少二次失真,并通过统计特征调整,保持数据的可用性。该方法在强化隐私保护性能的同时,实现水印信息盲提取,并具备极强的鲁棒性。实验证明,该方法有效降低了数据泄露和被盗用的风险,是一种同时实现敏感数据的隐私保护、泄密溯源与版权标识的集成式方案。A robust blind watermarking method for databases based on additive noise is proposed to address the issues of copyright infringement and privacy breaches in the process of big data sharing and publishing.This method integrates database watermarking with data desensitization techniques,leveraging the inherent distortion of additive Laplace noise.Watermark information is embedded through the ratio parity and noise modification mechanisms,significantly reducing secondary distortion.By adjusting statistical characteristics,the method preserves data usability.While enhancing privacy protection,it enables blind extraction of watermark information and demonstrates exceptional robustness.Experimental results show that this approach effectively reduces the risks of data leakage and unauthorized use,providing an integrated solution that simultaneously achieves sensitive data privacy protection,leakage tracing,and copyright marking.

关 键 词：数据库水印 数据脱敏 数据安全 隐私保护 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]