Comparing YOLO and U-net deep learning algorithms in chronic wound image segmentation.

For both datasets, paired t-tests showed highly significant gains ( p < 0.001) in IoU and DSC for the best YOLOv8 and YOLO11 models compared with U-Net, with Wilcoxon tests yielding consistent conclusions despite the small sample size ( n = 5).

also in 132,142 other papers