Classroom evaluation system based on face expression recognition
Aiming at the current problem of subjectivity and delay in classroom evaluation,based on the improvement of face expression recognition technology,explore the relationship between face expression and classroom evaluation,and construct a prac-ticable classroom evaluation model.With the help of neural network software PyTorch,MobileNetV2 is smoothly and tested with three different neural network models on FER2013 for comparative analysis.The results show that MobileNetV2 performs the best in the dataset in terms of combined accuracy and model size.After training the model on the MMAFEDB dataset to obtain a model with better accuracy,multiple modes were set up on the existing classroom evaluation system to correspond to different states dur-ing class.As a result,a new classroom evaluation model is provided.
万方数据
