C. ADHESIVE WBC SEGMENTATIONThis section will introduce different learning rates and controlparameters comparison. The advantages of U-Net allow itto be the basic network for comparison experiments withdifferent weight maps. The effects of data augmentation withdifferent networks are also shown. Finally, the proposedmethod is compared with other segmentation methods.For evaluation metrics, 4 measures were used in the experiment, namely Accuracy (acc), Intersection over Union (IOU)Score, F1 Score, and average F-measure. These metrics aredefined as in (6), (7), (10) and (11):