Development of an Automated YOLO-Based Digital Microscopy System for Leukemia Early Detection

Abstrak

This research explores the development of an automated classification system based on digital microscopy for the early detection of leukemia, utilizing the YOLO (You Only Look Once) object detection algorithm. This study addresses the challenges associated with manual blood smear analysis in clinical diagnostics. By leveraging a comprehensive dataset of hematological images, the YOLO algorithm was implemented to enable real-time detection and classification of blood cells, focusing on identifying pathological changes in cell morphology. The performance of the system was evaluated using key metrics such as accuracy, precision, recall, F1-score, and processing time. The results showed that the YOLO-based system achieved an accuracy of 92.5% in detecting abnormal blood cells, with notable strengths in precision and real-time processing capabilities. Although YOLO's accuracy is slightly lower, its superior recall and speed make it particularly advantageous for applications requiring rapid analysis. This study introduces an application of YOLO in digital microscopy, emphasizing the importance of balancing speed and accuracy in diagnostic tools, which is crucial for improving early detection and treatment outcomes for leukemia.

Identifier

• OpenAlex: https://openalex.org/W4417147406
• DOI: https://doi.org/10.48084/etasr.12790