AN INTELLIGENT COUNTERFEIT MEDICINE CLASSIFICATION PREDICTION SYSTEM USING MODIFIED YOLO: A SINGLE STAGE OBJECT DETECTOR

Abstract

Counterfeit medicines pose significant risks to public health, necessitating robust identification systems. This research is motivated by the requirement of healthcare whereas it outlines an intelligent counterfeit medicine prediction system using modified YOLOv11 (mod-YOLOv11), which is designed for single-stage object detection. The system improves YOLOv11 by incorporating an efficient backbone network and attention mechanism to improve feature extractors and classification performance. This work processes high-definition images of medicines to detect counterfeit products with minimal delay and hence it can be applied in real-time applications. The advanced features like adaptive spatial partitioning and efficient feature pyramid networks of YOLOv11 can detect counterfeits accurately even in adverse environments. The preprocessing of data can enhance precision and recall rates as well as the F1-score as compared to the standard object detection models. The system also applies lightweight architectures to minimize diversified computational complexity. Extensive experimentation achieves 83.23% accuracy and minimal time consumption of 234ms for 500 epochs. This research offers a tangible and feasible approach to the identification of counterfeits that helps the world combat counterfeit products, particularly fake drugs and safeguard the health of the population.