Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

On the (In)effectiveness of Mosaicing and Blurring as Tools for Document Redaction

On the (In)effectiveness of Mosaicing and Blurring as Tools for Document Redaction Abstract In many online communities, it is the norm to redact names and other sensitive text from posted screenshots. Sometimes solid bars are used; sometimes a blur or other image transform is used. We consider the effectiveness of two popular image transforms - mosaicing (also known as pixelization) and blurring - for redaction of text. Our main finding is that we can use a simple but powerful class of statistical models - so-called hidden Markov models (HMMs) - to recover both short and indefinitely long instances of redacted text. Our approach borrows on the success of HMMs for automatic speech recognition, where they are used to recover sequences of phonemes from utterances of speech. Here we use HMMs in an analogous way to recover sequences of characters from images of redacted text. We evaluate an implementation of our system against multiple typefaces, font sizes, grid sizes, pixel offsets, and levels of noise. We also decode numerous real-world examples of redacted text. We conclude that mosaicing and blurring, despite their widespread usage, are not viable approaches for text redaction. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Proceedings on Privacy Enhancing Technologies de Gruyter

On the (In)effectiveness of Mosaicing and Blurring as Tools for Document Redaction

Download PDF
15 pages

Loading next page...
 
/lp/de-gruyter/on-the-in-effectiveness-of-mosaicing-and-blurring-as-tools-for-MYaSKnBz5G
Publisher
de Gruyter
Copyright
Copyright © 2016 by the
ISSN
2299-0984
eISSN
2299-0984
DOI
10.1515/popets-2016-0047
Publisher site
See Article on Publisher Site

Abstract

Abstract In many online communities, it is the norm to redact names and other sensitive text from posted screenshots. Sometimes solid bars are used; sometimes a blur or other image transform is used. We consider the effectiveness of two popular image transforms - mosaicing (also known as pixelization) and blurring - for redaction of text. Our main finding is that we can use a simple but powerful class of statistical models - so-called hidden Markov models (HMMs) - to recover both short and indefinitely long instances of redacted text. Our approach borrows on the success of HMMs for automatic speech recognition, where they are used to recover sequences of phonemes from utterances of speech. Here we use HMMs in an analogous way to recover sequences of characters from images of redacted text. We evaluate an implementation of our system against multiple typefaces, font sizes, grid sizes, pixel offsets, and levels of noise. We also decode numerous real-world examples of redacted text. We conclude that mosaicing and blurring, despite their widespread usage, are not viable approaches for text redaction.

Journal

Proceedings on Privacy Enhancing Technologies – de Gruyter

Published: Oct 1, 2016

References

  • Some methods for classification and analysis of multivariate observations In Proceedings of the fifth Berkeley symposium on mathematical statistics and probability Oakland CA
    MacQueen
  • Quantifying information leakage in document redaction In Proceedings of the st ACM workshop on Hardcopy document processing
    Lopresti
  • Some methods for classification and analysis of multivariate observations In Proceedings of the fifth Berkeley symposium on mathematical statistics and probability Oakland CA
    MacQueen
  • Getting the face behind the squares : Reconstructing pixelized video streams In
    Cavedon
  • Video scrambling for privacy protection in video surveillance : recent results and validation framework In SPIE Defense Sensing for Optics and
    Dufaux
  • Visual privacy protection survey with
    Padilla
  • Keyboard acoustic emanations revisited on Information and System
    Zhuang
  • reconstruction of encrypted voip conversations on fon - iks In Privacy on
    White
  • sick cure for the evaluation of compositional distributional semantic models In Proceedings of the Ninth International Language Resources and Evaluation Conference
    Marelli
  • Why blurring sensitive information is a bad idea https dheera net projects blur
    Venkatraman
  • deblur in gradient domain Optical
    Chen
  • Preserving privacy by de - identifying face images Knowledge and Data Transactions on
    Newton
  • Who alerted the cia other secret secrets Rump session
    Naccache
  • An introduction to super - resolution text In Digital Document Processing
    Mancas
  • All your biases belong to us Breaking in In th
    Vanhoef
  • Visual privacy protection survey with
    Padilla
  • An empirical identification method of gaussian blur parameter for image deblurring Processing Transactions on
    Chen
  • All Your Base Are Belong to US In Proceedings of
    Ford
  • Residual information of redacted images hidden in the compression artifacts In
    Ho
  • tutorial on hidden markov models and selected applications in speech recognition of the
    Rabiner
  • Analysing and Exploiting the Mantin Biases in In Cryptology ePrint Archive Report
    Bricout
  • sick cure for the evaluation of compositional distributional semantic models In Proceedings of the Ninth International Language Resources and Evaluation Conference
    Marelli
  • reconstruction of encrypted voip conversations on fon - iks In Privacy on
    White
  • Preserving privacy by de - identifying face images Knowledge and Data Transactions on
    Newton
  • Who alerted the cia other secret secrets Rump session
    Naccache
  • Why blurring sensitive information is a bad idea https dheera net projects blur
    Venkatraman
  • tutorial on hidden markov models and selected applications in speech recognition of the
    Rabiner
  • Residual information of redacted images hidden in the compression artifacts In
    Ho
  • Analysing and Exploiting the Mantin Biases in In Cryptology ePrint Archive Report
    Bricout
  • deblur in gradient domain Optical
    Chen
  • Video scrambling for privacy protection in video surveillance : recent results and validation framework In SPIE Defense Sensing for Optics and
    Dufaux
  • All Your Base Are Belong to US In Proceedings of
    Ford
  • All your biases belong to us Breaking in In th
    Vanhoef
  • An introduction to super - resolution text In Digital Document Processing
    Mancas
  • Getting the face behind the squares : Reconstructing pixelized video streams In
    Cavedon
  • An empirical identification method of gaussian blur parameter for image deblurring Processing Transactions on
    Chen
  • Quantifying information leakage in document redaction In Proceedings of the st ACM workshop on Hardcopy document processing
    Lopresti
  • Keyboard acoustic emanations revisited on Information and System
    Zhuang