AI Empowered Quantitative Evaluation Method for Handwritten Chinese Character

In order to solve the problems of untimely evaluation, unspecific and ineffective feedback for writers and to improve writing quality in the daily standardized Chinese character writing practice of primary and secondary school students, a quantitative evaluation method for paper-pen handwritten standardized Chinese character based on neural network is proposed in this paper. It takes handwritten Chinese character by paper-pen writing as the evaluation object, and obtains the quantitative features of Chinese character through feature extraction of handwritten Chinese character image samples. On this basis, CNN classifier is used to complete the classification of handwritten Chinese character images. Then, based on the Gaussian distribution to fit the writing feature values of excellent samples, strict and loose normative interval thresholds are obtained. Finally, the deviation between the quantitative features of handwritten Chinese characters and the threshold is calculated, to realize the general quality evaluation and the detailed quantitative evaluation of strokes.