Design the first children’s health interactive system in the application market, which combines the functions of intelligent assisted TCM constitutions identification and tongue image identification. With the system, parents can have accurate understanding of the characteristics of children’s physiological and psychological development; adopt targeted nursing methods according to the results of constitution identification; and master the TCM care in family for children’s common diseases.
For the first time, Generative Adversarial Neural Network (GAN) was adopted to construct the intelligent identification model of children’s TCM constitutions. With .net as the development environment, and combining Mysql database, the B/S architecture Web application is developed.GAN, Decision Tree, MLP, SVM and other models were used for data simulation experiment comparison. GAN has the highest accuracy of 92.1%.
GAN is the optimal model for intelligentized identification of children's TCM physique, so that the system has the functions of sensitive recognition of children's tongue image, identification of children's TCM constitutions, inquiry of children's TCM recuperation of common diseases, etc.This system can be installed and applied in smart phones and tablet computers with an iOS or Android system, and WeChat, a social media platform can be used for science popularization and commercial operation for the system. Moreover, the application of Chinese-English bilingual interface is conducive to the international promotion of TCM. In January 2020, Shanghai Association of Traditional Medical Engineering issued the "Application Certificate of Results" for this system. In 2020, 2 papers related to this system have been published. I applied for one invention patent and obtained two software Copyrights. Relevant achievements have won 4 international awards.