The Research And Implementation Of Face Detection And Recognition

With the development of computer science, as a kind of biometric technology, face detection and recognition gets more and more attentions. Face detection and recognition has become one of the most active topics for computer vision and pattern recognition.In this thesis, a face detection and recognition system is developed under OpenCV with VC++2008 concerning face detection and recognition problems.According to the process of face detection and recognition, a series of techniques are involved which include color balance, light compensation, skin detection, Adaboost algorithm, dilation, erosion, smoothing algorithm and PCA algorithm.In this system, in order to assist face recognition, a method of updating the face database dynamically is adopted. The final test results turn out that the system can detect and recognize faces fast and precisely.The system is mainly composed of four parts:The first part is image preprocessing, considering that the image could be too bright or too dark and the iamges may be influenced by color deviation, color balance and light compensation are used to deal with the two problems respectively.The second part is skin detection, the picture processed is transformed into YCbCr color space due to the clustering features of skin under it, and by doing so,the skin area is separated while the rest pixels filled with green which could reduce the computing pressure in the following steps.In addition, dilation and erosion are used to erase the burrs and noises in the images.The third step is to detect faces in the image which is got from the procedures of the two steps above and the transition from YCbCr to gray with Adaboost algorithm.The final step is to recognize the test face with one-dimension PCA algorithm with the assistance of dynamical face database maintained by the system.After examined by the practical pictures,the system has shown satisfying results while detecting and recognizing frontal faces.