Automatic Impervious Surface Area Detection Using Image Texture Analysis and Neural Computing Models with Advanced Optimizers

Table 2

Testing performance.

Models	Statistics	Performance measurement indices
		CAR (%)	TP	TN	FP	FN	Precision	Recall	NPV	F1 score
GDM	Mean	93.389	424.700	415.800	24.850	34.650	0.944	0.930	0.922	0.935
	Std	3.428	23.046	41.248	15.203	35.373	0.035	0.055	0.081	0.027
Adam	Mean	97.050	435.550	437.900	19.150	7.400	0.958	0.983	0.983	0.970
	Std	0.522	11.182	11.388	4.542	3.216	0.010	0.007	0.007	0.005
Adamax	Mean	96.572	427.800	441.350	21.000	9.850	0.953	0.977	0.978	0.965
	Std	0.728	17.189	16.487	6.025	2.903	0.013	0.007	0.006	0.008
Nadam	Mean	97.311	427.250	448.550	17.450	6.750	0.961	0.984	0.985	0.972
	Std	0.420	12.980	11.923	4.444	3.064	0.010	0.007	0.006	0.005
AdamW	Mean	97.028	424.600	448.650	21.250	5.500	0.952	0.987	0.988	0.969
	Std	0.520	11.061	11.897	4.898	2.110	0.011	0.005	0.005	0.005
AMSGrad	Mean	96.556	435.800	433.200	18.500	12.500	0.959	0.972	0.972	0.966
	Std	0.840	12.659	14.020	4.955	5.005	0.011	0.011	0.011	0.008