Title :
Two-stage scheme for false alarm suppression on fast face detection
Author :
Hailiang Li ; Tao Cui ; Kangheng Wu ; Zhibin Lei
Author_Institution :
Dept. of Electron. & Inf. Eng., Hong Kong Polytech. Univ., Hong Kong, China
Abstract :
This paper describes a novel two-stage scheme for false alarm suppression on fast face detection which is mainly applied on real-time video applications, such as surveillance systems. The new scheme consists of the well-known Viola-Jones framework and face verifier based on eigenfaces for fast face detection while suppressing the false alarm to extremely low. The proposed scheme can be benefited from fusion of both local and global features. Experimental results show that our proposed method is able to overcome the inherent limitation of the Viola-Jones face detector without any compromise in speed.
Keywords :
eigenvalues and eigenfunctions; face recognition; feature extraction; video signal processing; video surveillance; Viola-Jones framework; eigenfaces; face verifier; false alarm suppression; fast face detection; local feature-global feature fusion; real-time video application; video surveillance; Detectors; Face; Face detection; Feature extraction; Principal component analysis; Support vector machines; Training; Adaboost; Eigenfaces; Face detection; Principal Component Analysis; Support Vector Machine; false alarm;
Conference_Titel :
Anti-counterfeiting, Security, and Identification (ASID), 2014 International Conference on
Print_ISBN :
978-1-4799-7117-6
DOI :
10.1109/ICASID.2014.7064970