Chinese Herbal Medicine Classification Using Convolutional Neural Network with Multiscale Images and Data Augmentation

Abstract

Correct use of Chinese herbal medicines is vital to life safety of the patients. Chinese herbal medicine classification is very important for the correct use of Chinese herbal medicines. Traditional methods like microfeature identification and physiochemical identification are inefficient due to the various kinds and different conditions of Chinese herbal medicines. Therefore, we adopt a multiscale convolutional neural network (CNN) model with data augmentation technology to classify Chinese herbal medicines. The data augmentation techniques solve the problem of less data on Chinese herbal medicines. Multiscale technology extracts more useful features for Chinese herbal medicine classification. The experiments show favorable accuracy.