基于人脸检测的模板匹配人脸跟踪算法研究

The research of face tracking algorithm with template matching based on face detection

提出了一种基于人脸检测的模板匹配人脸跟踪算法。该算法采用基于灰度的归一化互相关匹配算法实现人脸的实时跟踪;利用haar分类器人脸检测算法实现人脸的重定位和人脸方位检测,提高人脸跟踪的精确性和自适应性,同时也为基于灰度的归一化互相关匹配算法提供了有效的跟踪模板,减小算法的时间复杂度,提高算法的有效性。实验数据分析表明,该算法是实现了人脸的实时跟踪,且对于人脸遮蔽、姿态变换下的人脸方位跟踪和人脸恢复后的重定位具有很好的准确性和自适应性。

In this paper, a face tracking algorithm with template matching based on face detection is improved. Face tracking is implemented by NCC（Normalized Cross Correlation） matching algorithm base on image gray value. Haar classifier face detection algorithm is used for re-location of face and position detection of face, which improves the accuracy and adaptability for face tracking, and provides an effective tracking template of the NCC matching algorithm based on gray scale to reduce the time complexity of the algorithm and improve the effectiveness of the algorithm. The experimental data analysis shows that the algorithm realizes real-time face tracking, and has good accuracy and adaptability for face tracking and face restoration in the case of face mask and pose transformation.