This module contains a wrapper of the detector provided by the FFLD2 [1] [2] project. This module also provides the very powerful DPM model provided by [3].

The FFLD2 detector is also trainable.


[2]Dubout, Charles, and François Fleuret. “Exact acceleration of linear object detectors.” Computer Vision–ECCV 2012. Springer Berlin Heidelberg, 2012. 301-311.
[3]Mathias, Markus, et al. “Face detection without bells and whistles.” Computer Vision–ECCV 2014. Springer International Publishing, 2014. 720-735.