Based On The Arm9-specific Face Recognition System

Face Recognition means computer technology for identity recognition which analyzes and compares people's face characters. This technology is hot spots of Pattern Recognition and Artificial Intelligence. It refers to many interdisciplinary researches such as image processing, pattern recognition, physiology,etc. It's widely used in accessing control, elisor, electronic-commerce and video monitoring. The traditional high property automatic recognition system are most based on PC platform for on-line application,while the pratical application demands higher portability and accessibility. This paper researched and designed specific face recognition system based on ARM9.On the basis of investigation about the domestic and overseas present status of face recognition and embedded system, firstly designed the hardware system, selected the processor chip, camera transducer and related peripheral circles, built an embedded hardware platform with S3C2440 as core. Sencondly, transplanted the Linux kernel, developed drivers and application programs of USB,touch screen and camera. Thirdly, used QT as the tool to develop graphical user interface. In the respect of algorithm, designed face recognition based on Gabor and skin color, this algorithm adopt skin color as character for district detection and location, adopt Gabor for dimension reduction, in the end adopt Principle Component Analysis algorithm for feature extraction, adopt mix distance classifier to measure the feature matching, and then compared other algorithm with this one, checked the effect of this algorithm.Compared with the usual PC system, this system is more portable and accessible. Face recognition slgorithm based on Gabor and skin color has improves in speed and precision. This system is new application in embedded field, giving full play to advantages of embedded system.