A Pilot Study of Diabetes Mellitus Classification from rs-fMRI Data Using Convolutional Neural Networks

Table 2

Accuracies achieved from Inception-v4-Residual network across five runs.

Inception-v2-Residual	Training time per epoch (s)	Total time (m)	Top-1 accuracy (%)	AUC
run1	5.65	9.51	89.51	0.8708
run2	564	9.49	89.88	0.8862
run3	5.71	9.61	90.12	0.8721
run4	5.70	5.60	90.12	0.8762
run5	5.59	12.03	90.12	0.8391
# of params.	=0.09 M
Mean accuracy	89.95%
Mean AUC	0.8690