A randomly generated snowy footprint scene is tainted with noise
at 10% of the
signal power. The image is generated based upon the known
gait of each animal.
Each print at every possible orientation is convolved with
each
of the "interesting" areas specified in the previous step.
The data points for each animal are fit with a linear approximation.
Every print at every orientation is convolved with the first region
returned by the white space function. Probable track locations are
calculated
based on location, angle, and gait of the identified animal track.