The Implementation Of Eembeded Face Recognition System On The ARM

With the rapid development of the computer technology and biomechanical engineeringtechnology, biometric technology that recognizes person identity based on personalbiometric feature has been the first choice of the security field system. Comparing with thetraditional method of person identity recognition, the face biometrics has the advantagessuch as non-contact, good stability and difficult to imitate. So face recognition is a verybroad market prospects for the application of identity recognition technology. When the facebiometrics technology combined with the development of embedded face recognitionsystem has a small size, low cost, low power consumption and multiple functions and so on.This paper, based onf face biometrics in an embedded system to achieve specificproblems, mainly from the three aspects of the hardware platform selection, face detectionand face recognition algorithm simulation and embedded face recognition software designare designed in the realization of the embedded ARM platform. In hardware design, thehardware environment of the face biometrics system based on ARM processor wasintroduced, including the embedded processor、touch screen、camera and so on. In facedetection and recognition algorithms simulation part, fast face detection algorithm based onskin color and a stable and efficient PCA-LDA fusion algorithm for face recognition werestudied. Software design was developed by four modules. In the module of imageacquisition, the driver and application program of USB camera based on Linux weredesigned in order to implement image acquisition; in the module of face detection and facerecognition, the algorithms of the face detection and recognition through computersimulation were transplantated on the ARM; On this basis, graphical user interface based onQT of embedded face biometrics were designed and then the design of the whole systemwas completed. Finally, the experiment result of the whole system combined the softwarewith hardware shows that the development of the embedded face recognition system can runsuccessfully with the ARM platform and the performance is good.