Hotspot Detection with Machine Learning Based on Pixel-Based Feature Extraction

Table 5

Comparison with contest winners.

Case	Accuracy (%)				False alarm rate				Efficiency (%)				Time (min)
	1st	2nd	3rd	Ours	1st	2nd	3rd	Ours	1st	2nd	3rd	Ours	1st	2nd	3rd	Ours
B1	93.81	43.36	69.47	90.71	8.08	0.83	3.22	9.80	11.61	52.13	21.57	9.26	0.09	1.84	0.11	0.18
B2	98.00	21.64	67.54	69.54	40.85	1.10	11.78	11.54	2.40	19.71	5.73	6.03	8.20	23.68	6.17	13.99
B3	91.82	80.73	99.62	96.48	11.24	5.19	38.62	21.73	8.17	15.57	2.58	4.44	18.73	118.95	7.97	16.57
B4	83.85	64.58	79.17	75.52	19.41	4.98	69.65	6.85	4.32	12.97	1.14	11.03	1.26	21.97	1.71	3.68
B5	92.86	61.90	47.62	71.43	47.95	0.74	5.83	1.21	1.94	83.87	8.16	58.82	0.44	5.43	0.67	2.05
Avg.	1.16	0.67	0.90	1.00	9.44	0.35	3.62	1.00	0.78	3.00	0.82	1.00	0.55	5.49	0.46	1.00