| As a biological characteristic of the human body, human face is difficult to forge, never to lose, easy to carry and easy to use. Face recognition has been potentially used in security authentication system, records management and human-computer interaction. Most of traditional face recognition systems are based on PC platform, however, these must be portability and easy-to-operate in some practical applications. The development of embedded system makes it possible. In this paper, face detection and recognition is researched with embedded platform because of the wide application prospect of face recognition and embedded system.The hardware was built with JXARM9-2410 breadboard of ARM9. Firstly, embedded Linux platform was constructed from building cross development environment, porting Bootloader, porting Linux kernel, creating root file system to porting device driver. Then reliable and steady embedded software development environment was established for the subsequent application development. Secondly, Linux V4L was analized to accomplish image capture module; Face detection algorithm which was based on YCbCr color space achieved face detection program for the foundation of face recognition; Face recognition theory based on eigenface was applied to face recognition module. Finally,MiniGUI was used to finish the human-computer interface and function of this system. In our practice, some algorithms which have been optimized according to embedded features is more adaptable to embedded system. Experimental results show that the system not only successfully run on ARM platform but also has good performance. It can be usefull as guidance for the application and development of face detection and recognition in embedded platform. |
