Towards a privacy protection-capable noise fingerprinting for numerically aggregated data. Issue 119 (August 2022)

• Record Type: Journal Article
• Authors: Hu, Yun; Hu, Aiqun; Li, Chunguo; Li, Peng; Zhang, Chunyu
• Is Part Of: Computers & security. Issue 119(2022)
• Publication Date: 2022-08
• Subjects: Digital fingerprinting -- Differential privacy -- Traitor tracing -- Privacy protection -- Numerically aggregated data; Computer security -- Periodicals; Electronic data processing departments -- Security measures -- Periodicals; 005.805
• DOI: 10.1016/j.cose.2022.102755
• Languages: English
• ISSNs: 0167-4048

Abstract

Abstract: Privacy protection and traitor tracing have always been two separate research subjects. Meanwhile, digital fingerprinting, which tracks illegally distributed digital media by creating a unique code for each user, has long been regarded as one of the effective means of tracking down illegal users. However, most of the existing fingerprinting schemes focus on how to design a more robust or general fingerprint code to target illegal users more effectively, without paying attention to privacy protection itself. Our paper proposes a novel idea to simultaneously perform privacy protection and traitor tracing on massive aggregated data. And further, a detailed Secure Noise FingerPrinting based on Differential Privacy (SNFP-DP) scheme is proposed to demonstrate the feasibility of the idea by integrating differential privacy and digital fingerprinting. The SNFP-DP scheme builds the well-designed noise for each specific computing node into the numerically aggregated data set, realizing fingerprint inserting and noise disturbance in one step. The privacy of Noised Aggregated FingerPrint Data (NAFPD) generated by our SNFP-DP scheme can be protected by adding appropriate noise, while the source of illegal dissemination can be determined by identifying the fingerprint in the NAFPD. In addition, we instantiate the SNFP-DP scheme on real-world data by specifying the values of all key parameters. The experimental results show that the NAFPD meets the requirements of data privacy protection and statistical analysis accuracy. And it also has the robustness and security characteristics as the digital fingerprinting. Finally, we conduct a thorough performance analysis through mathematical formula derivation and security attack simulation.