more general.  Previously they required the user to use the same type of image
for both input and output.  This has been relaxed so now different typed images
can be used.

faster.

and doing interpolation at the same time.

tools they would probably prefer to use.

to #include <> syntax.

program lists

assignment problem learning tools to also include an optional
num_nonnegative_weights() method.  This method can be used to tell any machine
learning tools which elements of the learned parameter vector should be
non-negative.  As part of this change, I also removed the previous methods for
doing this from the structural_assignment_trainer since they are now redundant.

to cout.

unit tests.

of TCP connections in the TIME_WAIT state on Windows XP.  This results in the tests failing
if you run them multiple times within the TIME_WAIT timeout period.  So this change makes
the tests less likely to false alarm when run on Windows XP.

matrix with a sparse vector.