Does Appearance Influence Homework Rating?

Homework is a common way for schools to check teaching effectiveness. It is Observed that high-scoring homework is often rich in content and reasonable in structure, while low-scoring homework is often lack of content and messy in structure. In order to verify the observation that the appearance factors of homework have an impact on homework scores, this paper uses deep learning to analyze them. In this paper, the daily homework rated by teachers is collected, and the homework samples are categorized into four class: A, B, C and D, which are made up of Homework Data Set(HDS). The VGG16 and ResNet50 neural network models are trained with the data sets. Through visualization of the trained models, it is found that the appearance factors of students' homework have certain influence on homework evaluation.