基于人脸检测的人脸跟踪算法

文章提出了一种基于人脸检测技术的人脸跟踪算法。该算法利用前一帧的人脸检测结果预测当前帧中人脸可能的尺度与位置范围，在限定的范围内采用模板匹配与人工神经网分类的方法定位人脸，从而实现快速而可靠的人脸跟踪。由于使用了人脸检测技术，该方法可以自动定位初始人脸。实验表明该方法在具有复杂、动态变化背景的图象序列中是很有效的，速度为5－11Hz，可用于实时性系统。

An Algorithm for Tracking Faces Based on Face Detection

： In this paper，a face tracking algorithm based on face detection technique is presented．Fast and robust face tracking is achieved by locating faces using template matching and neural network classifier in the current frame under constrained searching space in both face size and its location that are predicted according to the detection results of the previous frame．Because of the face detection technique used，the algorithm can automatically locate initial faces．Experimental results demonstrate its effectiveness for tracking faces in image sequences with complex and dynamic changing background that achieves 5－11Hz in speed，which can be used in time critical systems．