基于亮瞳效应的连续AdaBoost人眼检测

主动红外下的人眼检测结合了亮瞳效应，能够大大提高人眼的检测速度。连续AdaBoost算法借助于积分图和级联结构，具有很快的运算速度，而且它具有检测率高和鲁棒性好的特点。结合了亮瞳效应的连续AdaBoost人眼检测，使得检测系统能够满足实时的应用场合。同时，论文中采用的“人脸-人眼”两层结构有效降低了人眼检测的误检率，提高了检测的精度。实验结果表明，对于640×480分辨率的连续视频，系统的人眼检测平均时间小于16ms每帧，检测正确率较高。

Eye detection based on real AdaBoost and bright pupil effect

Eye detection under active infrared combining bright pupil effect can greatly enhance the eye detection rate.With the help of the integral image and cascade structure,our real AdaBoost algorithm has fast computation speed and it also has a high detection rate and good robustness.Eye location combining bright pupil effect and real AdaBoost makes the system to meet the real-time applications.At the same time,the "face-eye" two layers structure used by this paper reduces the false detection rate effectively an...