YOLO-based Face Emotion Recognition System for Teaching Scenes

Abstract

The main way to get emotional data has been through how they looked, which are able to interact a variety of feelings like joy, grief, fear, rage, disgust, and surprise. These expressive emotions are vital to communication due to their enable people to better understand one another's thoughts and desires. Therefore, we thought that we could utilize technologies such as computer vision to obtain images of students' expressions in the classroom through cameras and other related devices. In order to create a database, the camera is used to build an image acquisition system. One image is preprocessed, the YOLOv5s algorithm is used to extract the features of the face, and the outcome of correct recognition is determined by comparing the feature vectors that were obtained. The expressions are then categorized using the extracted feature vectors based on a multilayer perceptual network (MLP). Lastly, the computer summarizes the identifying data and uses a WeChat applet to deliver it to the teacher.Teachers can more effectively comprehend the educational needs of their pupils surroundings and modify their approaches to instruction with the help of this WeChat applet. This will help improve the quality of online education.