Table 2: Performance evaluation of GGNs classification results using different composition of input image

	Accuracy(%)	Sensitivity(%)	Specificity(%)
Method A Overall accuracy 72.41%
GGN-P	75.00	66.67	90.00
PSN-S	80.77	71.43	84.21
PSN-L	84.00	100.00	80.95
Method B Overall accuracy 68.97%
GGN-P	71.43	72.22	70.00
PSN-S	76.92	85.71	73.68
PSN-L	83.33	25.00	95.00
Method C Overall accuracy 75.86%
GGN-P	81.48	83.33	77.78
PSN-S	81.48	57.14	90.00
PSN-L	84.62	75.00	86.36
Method D Overall accuracy 75.86%
GGN-P	81.48	83.33	77.78
PSN-S	75.86	42.86	86.36
PSN-L	91.67	100.00	90.00
Method E Overall accuracy 75.86%
GGN-P	81.48	83.33	77.78
PSN-S	84.62	42.86	100.00
PSN-L	81.48	100.00	78.26
Method F Overall accuracy 75.86%
GGN-P	75.86	77.78	72.73
PSN-S	81.48	57.14	90.00
PSN-L	91.67	100.00	90.00
Proposed Method Overall accuracy 82.76%
GGN-P	85.71	94.44	70.00
PSN-S	85.71	42.86	100.00
PSN-L	92.31	100.00	90.91