The labels of a model are now cloned instead of using Mat::copyTo, because Mat::copyTo leads to a crash with the Python wrapper. I need to further investigate it.

Issue #2231: Attempt to fix segfault, when copying labels. LDA needed to be updated to generate the Wrapper correctly.

facerec.cpp: Removed Original LBP wrapper, because it is not in use and therefore generates a warning.

facerec.cpp: Added an exception to the LBP extraction, so now a meaningful exception is thrown if the wrong image type was given. Thanks to Eric Christiansen for reporting.

FaceRecognizer supports updating a model now. Documentation has been updated to reflect latest changes.

Updated the train method to be an InputArrayOfArrays, so the Python wrapper now generates the correct code (and it works).

Fixed facerec_eigenfaces.cpp demo sample: it contained hardcoded values, which made wrong element access possible. Fixed unsigned integer/signed integer warnings in facerec.cpp.

Bug #2214: Fixed bug, when calling the train method multiple times. The Eigenfaces and Fisherfaces method now re-estimate the model, the LBPH method appends the new data to the existing model.

Documentation for FaceRecognizer started (not added to toc of contrib/doc/contrib.rst yet). CSV file for the samples/cpp/facerec_demo.cpp added. Added some more exceptions to warn the user, when passing bad aligned data.

Exceptions now go into CV_Error. Added thresholding to the FaceRecognizer and updated the demo accordingly.

Several exceptions added to the available FaceRecognizer classes and helper methods, so wrong input data is reported to the user. facerec_demo.cpp updated to latest cv::Algorithm changes and commented.

FaceRecognizer class is now derived from Algorithm, therefore it's possible to set and retrieve the parameters using conventional Algorithm::set and Algorithm::get methods

Suppressed hundreds of useless MSVC warnings (can be reenabled setting cmake variable ENABLE_NOISY_WARNINGS to ON). Fixed some of remaining warnings.