Face detection is one of the mostly studied problems in vision [31]. The seminal
work of Viola and Jones [26] has established the two foundation principles for
practical solutions: 1) boosted cascade structure; 2) simple features. Most (if
not all) realtime face detectors in academia and industry nowadays are based on
the two principles. Such detectors work well for near-frontal faces under normal
conditions but become less effective for faces non-frontal or under more wild conditions
(lighting, expression, occlusion), because the simple features like Harr in
the cascade training are insufficient to capture the more complex face variations