Revolutionizing Traditional Chinese Medicine Image Classification and Recognition with an Improved YOLOv5

With the continuous development and application of traditional Chinese medicine (TCM), automated recognition and classification of TCM herbs has become increasingly important. This thesis introduces a YOLOv5 model based on the GIOU loss function for the classification and recognition task of TCM images. The model combines the fast target detection capability of YOLOv5 and the superior performance of the GIOU loss function to achieve efficient and robust classification of TCM images. Experiments were conducted on a dataset containing a large number of TCM images to evaluate the performance of our model. The experimental results show that the improved YOLOv5 model has an average precision of 85.15% and an accuracy of 98.64% for a threshold range of 0.5 to 0.95, which is an improvement over the native yolov5s type. This meets the requirements of Chinese herbal medicine image classification in the actual application.