Improving Diagnostic Efficiency in Haematology: A YOLOv8-Driven Framework for Automated Leukemia and Anemia Detection

Timely and accurate diagnosis of hematological diseases like leukemia and anemia is crucial for prompt medical intervention and positive patient outcomes. The paper proposes an automated machine learning system based on the state-of-the-art YOLOv8 object detector to analyze microscopic blood smear samples. The training data used for the system are broad and well-selected and thus the system is stable to the variations in staining, image quality and cell morphology, common in clinical practice. The results obtained from the experimental activities demonstrate that this model is highly effective in the detection and classification of abnormalities of red and white blood cells. This YOLOv8 driven solution provides a fast, reliable fully automated diagnostic system that utilizes less time consuming manual analysis processes, human error is minimized and overall efficiency of the hematology laboratory is increased leading to faster and informed clinical decision making.