Quantifying information leakage in document redaction

ABSTRACT

In this paper, we examine ways in which sensitive information might leak through the process of redaction. Such attacks apply known methods from document image analysis and natural language processing to recover text thought to have been obliterated for the purposes of public release. Systematically identifying and testing these weaknesses is a first step towards designing effective countermeasures. We describe our development of a prototype semi-automated system intended to accept as input a redacted document and provide feedback to the user as to whether the document might suffer from such leaks.

REFERENCES

• 1: Adobe Systems Incorporated, San Jose, CA. Adobe Font Metrics File Format Specification, October 1998.
• 2: D. Butler. US intelligence exposed as student decodes Iraq memo. Nature, 429:116, May 2004.
• 3: C. Fang and J. J. Hull. A word-level deciphering algorithm for degraded document recognition. In Symposium on Document Analysis and Information Retrieval, pages 191--202, 1995.
• 4: J. J. Hull, S. Khoubyari, and T. K. Ho. Visual global context: Word image matching in a methodology for degraded text recognition. In International Conference on Pattern Recognition, pages 665--668, The Hague, Netherlands, 1992.
• 5: Siamak Khoubyari , Jonathan J. Hull, Font and function word identification in document recognition, Computer Vision and Image Understanding, v.63 n.1, p.66-74, Jan. 1996 [doi>10.1006/cviu.1996.0005]
• 6: G. E. Kopec. Least-squares font metric estimation from images. IEEE Transactions on Image Processing, 2(4):510--519, October 1993.
• 7: J. W. Leonard. Classified National Security Information Directive No. 1, September 2003. http://www.archives.gov/isoo/policy_documents/eo_12958_implementing_directive.html.
• 8: D. Lopresti and A. L. Spitz. Information leakage through document redaction: Attacks and countermeasures, July 2004. Submitted for publication.
• 9: D. Naccache, May 2004. Private communication.
• 10: U.S. Census Bureau: Name Files, September 2004. http://www.census.gov/genealogy/names/.
• 11: John K. Ousterhout, Tcl and the Tk toolkit, Addison-Wesley Longman Publishing Co., Inc., Boston, MA, 1994
• 12: U.S. Census Bureau: U.S. and World Population Clocks, September 2004. http://www.census.gov/main/www/popclock.html.
• 13: A. L. Spitz. Using character shape codes for word spotting in document images. In D. Dori and A. Bruckstein, editors, Shape, Structure and Pattern Recognition, pages 382--389. World Scientific, Singapore, 1995.
• 14: A. Lawrence Spitz, Progress in Document Reconstruction, Proceedings of the 16 th International Conference on Pattern Recognition (ICPR'02) Volume 1, p.10464, August 11-15, 2002
• 15: A. Lawrence Spitz , Paul Marks, Measuring the Robustness of Character Shape Coding, Selected Papers from the Third IAPR Workshop on Document Analysis Systems: Theory and Practice, p.1-12, November 04-06, 1998
• 16: Tcl Developer Xchange, September 2004. http://www.tcl.tk/.
• 17: YAWL (yet another word list), September 2004. http://metalab.unc.edu/pub/Linux/libs/yawl-0.3.tar.gz.
• 18: A. Zramdini and R. Ingold. Optical font recognition from projection profiles. Electronic Publishing, 6(3):249--260, 1993.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.7 DOCUMENT AND TEXT PROCESSING
      I.7.5 Document Capture
          Subjects: Document analysis

General Terms:
Algorithms, Measurement, Security

Keywords:
character shape coding, classified documents, document analysis, document redaction