Design of Intelligent classroom facial recognition based on Deep Learning
Design of Intelligent classroom facial recognition based on Deep Learning
复制标题
DOI：
10.1088/1742-6596/1168/2/022043
复制
发表时间：
2019-02
期刊：
影响因子：
--
通讯作者：
Jielong Tang;Xiaotian Zhou;Jiawei Zheng
中科院分区：
文献类型：
--
作者：
Jielong Tang;Xiaotian Zhou;Jiawei Zheng
智能教室（ITS）的出现是当前现代教学水平的一个重大进步。本文利用计算机视觉的目标识别技术设计了一个实时课堂评估系统。本文基于FER-2013人脸表情数据集和卷积神经网络（CNN），通过去除全连接层，利用深度可分离卷积结合剩余模块，建立了一种实时情感识别模型。实验结果表明，本文建立的模型具有较高的检测精度和鲁棒性，能够实现对学生课堂表现的实时评价，给教师快速反馈。
The emergence of Intelligent Classroom (ITS) is a significant improvement in the current level of modern teaching. This paper designs a real-time classroom assessment system that utilizes computer vision’s target recognition technology. Based on the FER-2013 facial expression dataset and convolutional neural network (CNN), this paper establishes a real-time emotion recognition model by removing the fully connected layer and using the depth separable convolution combined with the remaining modules. In the end, the model established in this paper shows high detection accuracy and robustness, and can realize real-time evaluation of students’ classroom performance to give teachers quick feedback.