Abstract

This paper describes a face detection framework that is capable of processing image or video fast while achieving high detection rate. There are three key constructions. First, the database of MIT is adopted and image examples are used to train a linear classifier, which is based on the Haar-like features. Secondly, in order to make the system efficiently we utilize Principal Component Analysis (PCA). Thirdly, considering the high false detection rate when only use Haar-like features, we take the edge contour detection as a compensation. Because face is a part of human body, a face-like image which does not belong to a human body can be detected using our technique. Experiment results show that our method has a low false detection rate.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call