基于数理形态学的近红外光图像实时人脸检测

该新型算法首先使用基于Harr特征和Adaboost算法的人脸检测方法寻找人脸候选区域,并对候选区域进行归一化处理;然后利用人眼瞳孔在近红外光图像中会生成白色光斑的特点,使用基于数理形态学的Quoit滤波器精确定位眼睛.为了减小不同瞳孔大小带来的影响,使用了多尺度的Quoit滤波器以提高准确性.实验表明,这种方法不仅准确性高而且速度快,达到了实时人脸检测的要求.

Real time near infrared light face detection based on morphology

This new face detection algorithm first used Harr feature and Adaboost algorithm for locating the face area, and then normalized the area into specified size. Utilizing the property of high reflection rate under near infrared light on pupil, Quoit filter based on morphology was used for eye detection. In order to deal with difference sizes of pupil, a multi scale filter was proposed for reducing both of false positive rate and false negative rate. The experience showed that this method is accuracy and fast. And it fits the requirement of real time face detection