DIRT is an implementation of Shai Avidan and Ariel
Shamir's "Seam Carving for Content Aware Image
Resizing" algorithms. It allows you to remove and
add seams to an image, resizing it in a
non-uniform manner. DIRT uses the filter
algorithms from DIIT to find the seams.
This release adds a new filter for the filterable
hiding algorithms, two new hiding algorithms, and
a couple of major interface improvements. The
interface now sports an embedding rate progress
bar, so you know in advance how much of the
available hiding space will be used. The interface
also has a new "Explain" button which gives a
small synopsis of how each algorithm works.
This release tunes the hiding algorithm "BattleSteg" and fixes the
interface bug which caused the algorithm options to come up incorrectly
when the window was opened for the second time.
Sample pairs analysis code has been implemented in
DIIT. With the pre-existing RS analysis, this tool
allows for accurate steganalysis of images.
Traditional (black and white) laplace graph
information can also now be produced by DIIT. The
bugs in some benchmarking formulas have been
corrected.
