Eugene Leitl [mailto:Eugene.Leitl@lrz.uni-muenchen.de] wrote,
> http://www.lrz-muenchen.de/~ui22204/out1.jpg
> http://www.lrz-muenchen.de/~ui22204/out2.jpg
The first file "out1.jpg" contains a half a meg of random numbers encrypted
into the picture, while the second file "out2.jpg" contains no hidden
message. They were created with outguess 0.2 or later, (or some similar
program) that attempts to keep the statistical analysis of the file
unchanged.
Both files have the same exact color distribution statistics:
grey: red: green: blue:
mean: 86.34 76.22 89.51 97.11
standard deviation: 58.40 68.92 56.18 45.27
median: 67 46 67 82
The encoding program obviously went to great pains to keep the basic
statistics unchanged.
However, the first file is larger than the second file. It defines more
colors than the second file, although its picture appears to be of lower
quality. The colors do not seem as good because there are more of them and
they cover a wider area of variation. Basically, the first picture is more
complicated and contains more data than is justified by the picture quality
and contents. The ratio of extraneous bits to actual required bits seems to
indicate a message size of about 25% of the total file size.
How did I do? Was I able to detect a file that had an "undetectable"
message in it?
-- Harvey Newstrom <www.HarveyNewstrom.com> Principal Security Consultant, Newstaff Inc. <www.Newstaff.com> Board of Directors, Extropy Institute <www.Extropy.org> Cofounder, Pro-Act <www.ProgressAction.org>
This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:40:57 MDT
