Table 2.

Performance insights by task.

Tasks	Models	Performances
Classification tasks	ResNet-50	98.37% accuracy (chest x-ray)
	DeiT-Small	92.16% accuracy (brain tumor)
	EfficientNet-B0	81.84% accuracy (skin cancer)
	Fine-Tuned ResNet50	98.20% accuracy, 99.00% precision, 98.82% recall, 98.91% F1-score (COVID-19)
Segmentation tasks	DenseNet-121 U-Net	0.79–0.87 precision, 0.92–0.97 recall
	Diffusion-CSPAM-U-Net	84.4% DSC, 73.1% IoU
	Attention UNet	85.36% IoU, 91.49% Dice score
Detection tasks	YOLOv5-v8	95%–99.17% precision, 97.5% sensitivity, >95% mAP
	YOLO-NeuroBoost	99.48% mAP (brain tumors)
	YOLOv10	20 ms inference time (kidney stones)