c g a r c i a P i r i s a . f r Abstract. We present a novel face detection approach based on a convolutional neural architecture, designed to d e t e c t and precisely localize highly variable face p a t t e r n s , i n complex real world images. Our system automatically synthesizes simple problem-specific feature e x t r a c t o r s f r o m a training set of face and n o n face patterns, without m a k i n g a n y assumptions or using a n y hand-made design concerning the features to extract

more »

... the a r e a s of the face pattern to analyze. E x p e r i m e n t s on different difficult test sets have shown that our approach provide superior overall detection results, while being computationnally more efficient than most of state-of-the-art approaches that require dense scanning and local preprocessing. 0-7803-8178-5/03/$17.00 0 2003 IEEE