Detection, classification, and counting blood cells using YOLOv8

Abstract

Distinguishing normal white blood cells from leukemia cells plays a role in assisting in the diagnosis of blood diseases. Up to now, many automated methods are applied to provide more time efficiency, timing, and accuracy such as YOLO, SVM, CNN, and Faster CNN. In this research, we propose to use YOLOv8 to detect, classify, and count normal white blood cells and leukemia cells. The result of the experiment method with accuracy is 95.1% using the first number of blood images is 1500 images from AML public source and Hanoi Medical University collected by ourselves, and after augmentation is 3629 images that are used for training, validation, and testing. At the same time, we compare the accuracy of this method with studies using different versions of YOLO.