The Face Tracking Research Based on Cam-Shift and Improved NMI

In order to take the complex background interference and real-time requirements into account, the experience apply NMI feature and Cam-Shift feature space on face tracking. By secondary search on the feature space of NMI human eye´s feature and Cam-Shift emplacement space, the experience improve the accuracy of face identified location, adding Upper and lower line for improving calculate speed of characteristics of the human eye in NMI feature space leads to improve the accuracy of face detection. Experiments show that the method is real-time and effective.