WCSE

Evaluating YOLO Models for Enhanced Safety in Medication Dispensing

Abstract— This study assesses the effectiveness of YOLOv5 and YOLOv8 in enhancing safety in medicationdispensing through precise identification of capsules and tablets. Using a dataset of 1,659 images, the models were evaluated across various metrics in training, validation, and testing phases. Results indicate that YOLOv8 outperforms YOLOv5 in most training and validation metrics, while YOLOv5 shows superior performance in testing. These findings highlight the potential of advanced object detection models to improve patient safety by reducing medication dispensing errors, offering valuable insights for the deployment of AI in healthcareenvironments.

Index Terms— YOLOv5, YOLOv8, Pharmaceutical object detection, Deep learning in healthcare

Anan Panphuech
Suranaree University of Technology, THAILAND
Khwanchai Huailuk
Suranaree University of Technology, THAILAND
Sayan Kaennakham
Suranaree University of Technology, THAILAND

[Download]

Cite: Anan Panphuech, Khwanchai Huailuk, Sayan Kaennakham, "Evaluating YOLO Models for Enhanced Safety in Medication Dispensing," 2024 The 14th International Workshop on Computer Science and Engineering (WCSE 2024), pp. 81-85, Phuket Island, Thailand, June 19-21, 2024.