An English Handwriting Quality Evaluation Algorithm Based on Machine Learning

English handwriting assessment is an important part of basic English teaching. Started in 2001 in the United States, the International English Calligraphy Contest is a worldwide non-profit event sponsored by the International English Calligraphy Contest Association. Due to the large number of participants, the evaluation of English calligraphy is labor-intensive, so we would like to have English calligraphy evaluation algorithms that can evaluate English calligraphy and thus reduce the burden of judges to read the scripts. In the context of the epidemic, students need to complete the competition at home and upload photos using their cell phones. Therefore, in this paper, we propose an algorithm to automatically evaluate the quality of English handwriting, hoping to achieve the function of correcting the test papers uploaded by cell phones and giving scoring results. In this paper, we use traditional image processing to calibrate the photos of the test papers taken by cell phones and combine traditional feature extraction methods with convolutional neural networks to jointly extract the features of English handwriting. Our algorithm not only enables us to achieve 99% accuracy in our experiments, but also improves the interpretability of the English handwriting evaluation algorithm to some extent.