Abstract
There are four methods for diagnosing diseases in TCM: “looking, smelling, asking and cutting”, and tongue diagnosis is an indispensable part of TCM diagnosis. According to Chinese medicine, the tongue has a close relationship with the internal organs, meridians, qi and blood, and fluids. However, traditional Chinese medicine tongue diagnosis requires the clinical experience of doctors and the surrounding environment, and the human eye has limited recognition of the same tongue image, so there is a lack of objective, quantitative and standard evaluation and application. The TCM pediatric tongue diagnosis model supported by AI deep neural network technology can provide qualitative and quantitative feature evaluation indexes, making it possible to reduce the error of diagnosis. The problem of clinical data scarcity is faced by all, and the construction of data simulation enhancement models and data classification and discrimination models applicable to small data sets are innovative and meaningful work.