基于面部特征的线上关注度检测与判断

Online attention detection and judgement based on facial features

针对线上教学时学生注意力易分散、关注度不高的问题,设计基于面部特征的线上关注度检测与判断的方法。首先,利用Dlib库检测视频图像中的人脸和眼部特征点;其次,对左右眼睛的2个外眼角不在同一水平线上的人脸图像进行矫正,在矫正后的人脸图像中根据眼部特征点的位置获取眼部图像,通过眼部图像预处理确定虹膜和瞳孔区域,进而求出瞳孔中心的位置;最后,根据瞳孔中心在眼部图像中的位置判断眼睛是否向左或向右注视,分析眼睛是否关注屏幕。实验结果表明,在人脸无任何遮挡的状态下,所设计的方法能较好地定位瞳孔中心位置并分析眼部注意力状态,对学生线上关注度做出判断。

In order to solve the problem that the attention of students is easy to be distracted and their attention decreases in online teaching,an online attention detection and judgment method based on facial features is designed.Firstly,the Dlib library is used to detect the face and eye feature points in the video image.Secondly,the face image with two external corners of the left and right eyes not on the same horizontal line is corrected.In the corrected face image,the eye images are obtained according to the positions of eye feature points,the iris and pupil areas are determined by eye image preprocessing,and then the positions of the pupil centers are obtained.Finally,the positions of the pupil centers in the eye images are used to determine whether the eyes look left or right,and analyze whether the eyes pay attention to the screen.The experimental results show that the positions of the pupil centers can be located well and the eye attention state can be analyzed when the face is free of any occlusion,and the online attention of students can be judged.