Fix the documentation to use bob.ip.facedetect instead of opencv

See merge request !12

comment opencv instructions from docs

opencv is not very stable and is not a dependency of this package anymore.
Testing it in doctests is causing this package to fail which should not be the case.

See merge request !11

Resolve "Full image does not work in locate()"

Closes #7 

I have implemented the default values for the bounding box and solved the issue that happened when the (automatically extended) bounding box was outside the image boundaries.

See merge request !10

Added default values for y,x,height,width in Flandmark.locate(); added test to assure that full image resolution works

[Automatic] update links and the ci mostly



See merge request !9

Remove OpenCV dependence

Added bounding box offset, which seems to be required to make the results compatible with OpenCV image processing

This reverts commit 8063f3ec.