Predicting the Content of the Main Components of Gardeniae Fructus Praeparatus Based on Deep Learning

Gardeniae Fructus (GF) and its stir-fried product, Gardeniae Fructus Praeparatus (GFP), are commonly used herbal medicines in traditional Chinese clinic. However, it is challenging to measure the content of GFP’s main components rapidly during processing. In this paper, an MLP-based method for GFP component content prediction is proposed. 10 deep learning models including CNN and Transformer are used to extract features from the built image dataset. Combined with the measured component content data, the extracted feature data are used to train the MLP regression model and evaluate the effect. It is demonstrated that the proposed method can be used for the rapid and nondestructive determination of the content of the main components of Chinese herbal pieces. This study provides insights for similar studies.