non-square matrix.

out by Yan Xu.  One of the face locations was mistakenly used twice while
another was skipped.  This change breaks backwards compatibility with the
previous feature extraction output but should slightly improve accuracy of
classifiers trained using these features.

switching to projected gradient steps.

Revise help message

and try to link against that one rather than whatever might be installed
system wide.

installed already.

some OS X machines when load_libsvm_formatted_data() was called.

Add a Python interface for the Correlation Tracker

This replicates the c++ example.

This is useful for visualizing the tracking results.

Added the missing documentation for the correlation tracker.
Didn't realise that the tracker returned the side lobe ratio,
so fixed the API to return it properly.

This aids an API for the correlation based tracker that Davis
recently added to Dlib. I've made sure to allow overrides for
passing normal rectangles rather than drectangles (which isn't
currently supported in the Dlib C++ API). This is mostly
because I imagine people might initialize the tracking using
something like a bounding box from a detector (e.g.
load_frontal_face_detector).

The correlation tracker deals with the drectangle
(double rectangle) class which currently isn't wrapped. Therefore,
I add the drectangle class and refactor rectangles into their
own file. I also added a load of methods on rectangle that might
be useful such as intersection/contains/area etc.

The word detector didn't make much sense!

learning tools.