Multi-view Face Detection under Unconstrained condition

Multi-view face detection is the main problem we consider in this paper. There has been significant research on this problem, while current state-of-the-art algorithms for this task leave something to be desired. In this paper, we present an algorithm to enhance the face detection performance. Before images were put into the deep belief nets (DBN) model, they were preprocessed in turn, and then capitalized on DBN to train the data automatically. Use FloatBoost (FB) Algorithm to learn these feature, and classify them based on pyramid architecture. Evaluations on dataset show that our proposed algorithm has similar or better performance compared to the current methods. It can accurately spot faces at multi-view, even when partially occluded.