Table 6.

The pre-trained models performance in six categories on the difference image format.

Category = 6 (Healthy and Parkinson Disease Stages I to V.)
Net	Image Format	Batchsize	MaxEpochs	Accuracy	Recall	Precision	F-Score	Kappa
AlexNet	Gray	96	48	0.774	0.742	0.853	0.794	0.679
	Color	120	60	0.521	0.349	0.491	0.408	0.282
GoogLeNet	Gray	34	17	0.617	0.532	0.695	0.603	0.439
	Color	28	14	0.556	0.378	0.478	0.422	0.362
VGG19	Gray	24	12	0.778	0.612	0.665	0.637	0.683
	Color	22	11	0.754	0.746	0.747	0.747	0.669
ResNet50	Gray	20	10	0.681	0.56	0.768	0.647	0.537
	Color	18	9	0.754	0.661	0.758	0.706	0.651
ResNet101	Gray	16	8	0.722	0.602	0.645	0.623	0.603
	Color	18	9	0.770	0.699	0.777	0.736	0.673
DenseNet201	Gray	16	8	0.762	0.670	0.739	0.703	0.661
	Color	8	4	0.778	0.696	0.814	0.750	0.680