Deformable Part Models are top performers at the Pascal VOC challenge, but they are quite slow (few secs per image on a strong machine). Usually they use HOG + SVM on a multi-scale search-space.

Papers+code: http://www.cs.berkeley.edu/~rbg/latent/

Another implementation with pre-trains models:

http://www.uco.es/~in1majim/proyectos/libpabod/

Aggregate Channel Features on the other hand are quite fast (30+ fps on a single core, depending on the implementation), since they just estimate the features on nearby scales:

Paper: http://research.microsoft.com/pubs/220572/DollarPAMI14pyramids.pdf

Code: http://vision.ucsd.edu/~pdollar/toolbox/doc/index.html

Mixtures of Parts or Mixtures of Tree is similar to Deformable Part Models could be a good reference for estimation of body and face poses: 

For human body:

http://www.ics.uci.edu/~dramanan/papers/pose2011.pdf

It is applied well to unique framework of three tasks of face deetection, pose estimation, landmark detection.

http://www.ics.uci.edu/~xzhu/face/

Take a look at the work by  Rodrigo Benenson http://rodrigob.github.io/

Especially at his paper:

M. Mathias, R. Benenson, M. Pedersoli, L. Van Gool:  Face detection without bells and whistles. ECCV 2014.

You can get the detector used in the paper at http://markusmathias.bitbucket.org/2014_eccv_face_detection/