Research on Calligraphy Evaluation Technology Based on Deep Learning
Zhang, Zhi (2023-05-10)
Research on Calligraphy Evaluation Technology Based on Deep Learning
(10.05.2023)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
avoin
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe2023060552563
https://urn.fi/URN:NBN:fi-fe2023060552563
Tiivistelmä
Today, when computer-assisted instruction (CAI) is booming, related research in the field of calligraphy education still hasn’t much progress. This main research for the calligraphy beginners to evaluate their works anytime and anywhere. Author uses the literature research and interview to understand the common writing problems of beginners. Then conducts discussion on these problems, design of solutions, research on algorithms, and experimental verification. Based on the ResNet-50 model, through WeChat applet implements for beginners. The main research contents are as follows:
(1) In order to achieve good results in calligraphy judgment, this article uses the ResNet-50 model to judge calligraphy. First, adjust the area of the handwritten calligraphy image as the input of the network to a small block suitable for the network. While training the network, adjust the learning rate, the number of image layers and the number of training samples to achieve the optimal. The research results show that ResNet has certain practicality and reference value in the field of calligraphy judgment. Regarding the possible over-fitting problem, this article proposes to improve the accuracy of the judgment by collecting more data and optimizing the data washing process.
(2) Combining the rise of WeChat applets, in view of the current WeChat applet learning platform development process and the problem of fewer functional modules, this paper uses cloud development functions to develop a calligraphy learning platform based on WeChat applets. While simplifying the development process, it ensures that the functional modules of the platform meet the needs of teachers and beginners, it has certain practicality and commercial value. After the development of the calligraphy learning applet is completed, it will be submitted for official.
(1) In order to achieve good results in calligraphy judgment, this article uses the ResNet-50 model to judge calligraphy. First, adjust the area of the handwritten calligraphy image as the input of the network to a small block suitable for the network. While training the network, adjust the learning rate, the number of image layers and the number of training samples to achieve the optimal. The research results show that ResNet has certain practicality and reference value in the field of calligraphy judgment. Regarding the possible over-fitting problem, this article proposes to improve the accuracy of the judgment by collecting more data and optimizing the data washing process.
(2) Combining the rise of WeChat applets, in view of the current WeChat applet learning platform development process and the problem of fewer functional modules, this paper uses cloud development functions to develop a calligraphy learning platform based on WeChat applets. While simplifying the development process, it ensures that the functional modules of the platform meet the needs of teachers and beginners, it has certain practicality and commercial value. After the development of the calligraphy learning applet is completed, it will be submitted for official.