Enhancing Chinese Font Recognition with Convolutional Neural Networks

With more and more styles of fonts, young people are keen to use distinctive fonts to show their individuality. The purpose of this paper is to study Chinese font classification recognition to enable people to handle fonts more effectively. For this purpose, Nankai Chinese font style dataset is used for subsequent training and testing. The dataset was processed by normalizing the central moment features. The data was imported into the Sequential model of keras, and the accuracy of font classification was obtained after convolutional processing as 94.67%. The accuracy and running speed of the font recognition model in this study are excellent. The results show that this model is able to solve problem font recognition outstandingly after learning and training.