Chinese Herbal Medicine Leaves Classification Based on Improved AlexNet Convolutional Neural Network

摘要

In recent years, deep learning has been widely used in various fields in image processing. In this paper, an improved AlexNet model is proposed to enhance feature extraction, and its feasibility is verified by comparative experiments. Firstly, the web obtainer algorithm is used to climb the image of 5 types of Chinese herbal medicine leaves to form a data set with a small sample size. Secondly, the original data set capacity is increased by 4 times to form a new data set by data augmentation technology. Finally, four sets of comparative experiments are carried out between the improved AlexNet model and the augmented data set. The experimental results show that the AlexNet model combined with data augmentation and improvement can greatly improve the accuracy of Chinese herbal medicine image classification.