JPEG-Based Microdata Protection

  • Javier Jiménez
  • Guillermo Navarro-Arribas
  • Vicenç Torra
Conference paper

DOI: 10.1007/978-3-319-11257-2_10

Volume 8744 of the book series Lecture Notes in Computer Science (LNCS)
Cite this paper as:
Jiménez J., Navarro-Arribas G., Torra V. (2014) JPEG-Based Microdata Protection. In: Domingo-Ferrer J. (eds) Privacy in Statistical Databases. PSD 2014. Lecture Notes in Computer Science, vol 8744. Springer, Cham

Abstract

JPEG-based protections can be obtained by regarding microdata as an image that is transformed by means of a lossy JPEG compression-decompression process. Here we propose a general model that decouples JPEG-based methods into two parts. First part encompasses transformations between data and image spaces. Second part consists in the image transformation itself. Under this general model, we first explore different maps between data and image spaces. In our experiments, quantization using histogram equalization, in combination with JPEG-based methods, outperform other approaches. Secondly, image transformations other than JPEG can be utilized. We illustrate this point by introducing JPEG 2000 as a valid alternative to JPEG. Finally, we experimentally analyze the effectiveness of the generalized JPEG-based method, comparing it with well-known state-of-the-art protection methods such as rank swapping, microaggregation and noise addition.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Javier Jiménez
    • 1
  • Guillermo Navarro-Arribas
    • 2
  • Vicenç Torra
    • 1
  1. 1.IIIA-CSIC, Artificial Intelligence Research InstituteSpanish National Research CouncilSpain
  2. 2.DEIC-UAB, Dep. of Information Engineering and CommunicationsUniversitat Autònoma de BarcelonaSpain