【摘要】 人脸表情是表达人类感情最重要的方式之一。近年来,人脸表情识别在人机交互、公共安全监控、教育、医疗、交通等方面得到越来越多的应用。本文深入探讨了人脸表情识别在辅助医疗中的应用,以及提出了一种使用卷积神经网络MTCNN进行人脸检测和对齐,使用卷积神经网络Xception进行人脸表情识别的方法。经实验证明,这两种卷积神经网络在保证准确率不低于传统方法的情况下,大大优化了网络结构和提高了速度,为本实验中构建一个轻量级的实时人脸表情识别网络模型奠定了基础。更多还原

【Abstract】 Facial expressions are one of the most important ways humans express feelings. In recent years, facial expression recognition has been widely used in human-computer interaction, public safety monitoring, education, medical care,transportation, etc. This paper discusses in detail the application of facial expression recognition to assistive medical care and proposes a method for real-time facial expression recognition using two lightweight convolutional neural networks(CNNs).Experiments prove that these two CNNs greatly optimize the network structure and improve speed, while ensuring that the accuracy is comparable to or better than the traditional method, and lay the foundation for real-time facial expression recognition.更多还原