自动人脸跟踪方法研究

Research of automatic face tracking

针对实时视频监控领域中传统的Camshift算法不能自动跟踪人脸和容易受到肤色相近遮挡等问题,首先采用Adaboost算法实现了人脸的自动检测,同时对于跟踪丢失等情形,通过贪心预测和卡尔曼预测对跟踪偏差进行实时改进,并比较两种算法的优缺点。实验表明前者对跟踪的准确性有较大提高,后者具有较好的实时性,在相近肤色遮挡时仍能实现正确跟踪,并对侧脸也有较好的效果,算法具有较好的鲁棒性。

In the field of real-time video monitoring,the classic Camshift algorithm can not automatically track the human face and easily subject to color interference covered and other issues,the paper first used Adaboost algorithm to detect the human face automatically,and also for the situation when the track was lost,through the greed forecasting and Kalman prediction to improve the tracking error in real-time,compared the advantages and disadvantages of the two algorithms. Experimental results show that the accuracy of the former on the track has improved greatly,the latter has better real-time,even when the block in the similar color to achieve the right track,and side faces also have a good effect,and the algorithm has better robustness.