Automated Blood Cell Detection and Classification in Microscopic Images Using YOLOv11 and Optimized Weights

Area of Science:

• Medical Imaging
• Hematology
• Computer Vision

Background:

• Accurate blood cell detection is vital for diagnosing hematological disorders.
• Automating this process can significantly improve diagnostic efficiency.

Purpose of the Study:

• To develop and evaluate advanced YOLOv10 and YOLOv11 models for automated blood cell detection and classification.
• To focus on identifying red blood cells (RBCs), white blood cells (WBCs), and platelets for complete blood count (CBC) analysis.

Main Methods:

• Utilized the Blood Cell Count Detection (BCCD) dataset, enhanced with data augmentation.
• Conducted experiments with complete weight initialization and advanced optimization for YOLOv11.
• Performed meticulous hyperparameter tuning for the YOLOv11 architecture.

Main Results:

• The YOLOv11-l model achieved a mean Average Precision (mAP) of 93.8%.
• Demonstrated robust accuracy in classifying multiple blood cell types.

Conclusions:

• The YOLOv11 architecture is highly effective for automated blood cell classification.
• This technology shows significant potential for improving hematological analyses and clinical diagnosis.