人脸识别论文：A novel approach to access control based on face recognition.pdfVIP

A NOVEL APPROACH TO ACCESS CONTROL BASED ON FACE RECOGNITION A. Hadid, M. Heikkila, T. Ahonen, and M. Pietikainen ¨ ¨ Machine Vision Group Infotech Oulu and Department of Electrical and Information Engineering P.O. Box 4500 FIN-90014 University of Oulu, Finland ABSTRACT to access control suffer from either low recognition speed In this paper, we introduce an approach to access control due to the complex analysis of facial images, or to the low based on detecting and recognizing faces with novel meth- recognition accuracy. The main methods employed by fa- ods. The goal is to build an automatic face recognition sys- cial recognition vendors to identify and verify faces include tem which authorizes the access to the laboratory room for eigenfaces [3], Elastic Bunch Graph Matching [4], and Lo- only certain persons while denying the access for the others. cal Feature Analysis (LFA) [5] which is claimed to be used For this purpose, a camera is set in the door looking at the in Visionics commercial system FaceIt. A list of some com- corridor and the captured frames are analyzed. The analysis mercial systems is given in [1] and their performances are starts by selecting the regions of interest in the images us- reported in the Face Recognition Vendor Test (FRVT) [2]. ing either texture-based background subtraction or skin lo- Recently, we have proposed a new approach for face cus based color segmentation. Once the regions of interest recognition based on local binary patterns (LBP) [6]. Our are defined, a fast face detection scheme is launched. Then, method clearl