Artificial intelligence assists identification and pathologic classification of glomerular lesions in patients with diabetic nephropathy

Table 2 Performance of the V-Net architecture from the ARPS system on identifying three glomerular intrinsic cells in patients with DN

Intrinsic cells	Accuracy	Specificity	Precision	Recall	F1-score
Mesangial cells	0.942	0.953	0.932	0.926	0.929
Endothelial cells	0.930	0.970	0.939	0.855	0.895
Podocytes	0.961	0.986	0.890	0.793	0.839

ISSN: 1479-5876