An Enhanced YOLOv5 Algorithm with Coordinate Attention for Efficient Tuberculosis Bacteria Detection

Efficient and accurate diagnosis of Mycobacterium tuberculosis is crucial for tuberculosis control. However, traditional diagnostic methods rely on laboratory equipment, making them costly and time-consuming. To improve detection performance, a refined YOLOv5 framework embedding coordinate attention is introduced to strengthen feature representation in challenging environments. A large-scale image dataset of tuberculosis bacilli served as the foundation for training, validation, and evaluation. Model effectiveness was evaluated using precision, recall, and mean average precision at IoU thresholds of 0.5 and 0.5:0.95. Compared to the original YOLOv5 architecture, the enhanced version demonstrated notable improvements, it achieved a 2% increase in precision, 3% improvement in recall, 3.3% boost in mAP@0.5, and 1% enhancement in mAP@0.5:0.95. These advancements were achieved without compromising the model's real-time detection capabilities. The proposed method significantly improves detection accuracy and computational efficiency, providing a robust AI-driven solution for medical image analysis. It has the potential to be integrated into automated screening systems, facilitating early tuberculosis diagnosis and enhancing disease control efforts.