Haar-like Features for Robust Real-Time Face Recognition

Face recognition is still a very challenging task when the inputface image is noisy, occluded by some obstacles, of verylow-resolution, not facing the camera, and not properly illuminated.These problems make the feature extraction andconsequently the face recognition system unstable. The proposedsystem in this paper introduces the novel idea of usingHaar-like features, which have commonly been used forobject detection, along with a probabilistic classifier for facerecognition. The proposed system is simple, real-time, effectiveand robust against most of the mentioned problems. Experimentalresults on public databases show that the proposedsystem indeed outperforms the state-of-the-art face recognitionsystems