An Intelligent Recognition System for Chinese Medicinal Materials Based on YOLOv8

The image recognition of traditional Chinese medicinal materials represents a significant direction in the informatization and intelligent development of traditional Chinese medicine, showcasing vast research and application prospects. This paper focuses on the diverse types of medicinal materials and the subtle differences in their images, designing and implementing an intelligent recognition system based on the YOLOv8 model. The overall architecture of this system adopts a modular design, integrating core functionalities such as image importation, model inference, result visualization, information recording, and support for multi-source input. The interface is developed using PyQt5, ensuring user-friendly interaction and ease of operation. During the model training phase, we constructed an image dataset specifically for traditional Chinese medicinal materials and optimized YOLOv8 through transfer learning. In functional testing processes, this system demonstrated excellent recognition accuracy and operational stability under various input conditions; achieving an average precision rate of $98.4 \%$ with an average inference time of 0.045 seconds, thereby possessing real-time application capabilities.