疲劳检测

说明：  虽然 AdaBoost 算法的检测速度快，误识率低，但是在样本的权重更新过程中，对于分类错误样本中的正、负样本没有加以区分，不利于提高正样本的识别率。本文提出一种新的权重更新方法，对于分类错误的样本，对判断错误的正样本给更高的权重，使得算法在下一轮迭代时，更加关注对分类错误的正样本的学习，从而提高对正样本（人眼）的检测率。采用基于最小二乘法对眼部的外轮廓进行椭圆拟合，根据拟合椭圆的参数来判断眼睛的睁闭状态；采用结合 PERCLOS 和眨眼频率的方法，对疲劳状态进行检测。(Although AdaBoost algorithm has fast detection speed and low false recognition rate, in the process of sample weight updating, there is no distinction between positive and negative samples in classification error samples, which is not conducive to improving the recognition rate of positive samples. In this paper, a new weight updating method is proposed, which gives higher weight to the positive samples of classification errors, so that the algorithm pays more attention to the learning of the positive samples of classification errors in the next iteration, so as to improve the detection rate of the positive samples (human eyes). Based on the least square method, the eye contour is fitted with ellipse, and the opening and closing state of eyes is judged according to the parameters of fitting ellipse; the fatigue state is detected by combining PERCLOS and blinking frequency..)

疲劳检测, 0 , 2020-06-27
疲劳检测\GUI.fig, 18622 , 2020-02-04
疲劳检测\GUI.m, 8884 , 2020-02-04
疲劳检测\GUI界面.png, 117846 , 2020-02-04
疲劳检测\GUI运行图.png, 315763 , 2020-02-04
疲劳检测\GetAllDatabase.m, 133 , 2018-05-05
疲劳检测\GetEyeAccurateDatabase.m, 981 , 2018-05-05
疲劳检测\GetEyeAccurateImg.m, 1483 , 2008-05-03
疲劳检测\GetFaceDatabase.m, 698 , 2018-05-05
疲劳检测\GetValideDatabase.m, 720 , 2007-05-04
疲劳检测\GetValideImage.m, 809 , 2000-05-04
疲劳检测\bwe.mat, 53232 , 2008-05-03