Table 4.

Classification results of five image features using different classifiers in the 120 × 120 pixels sub-database of GasHisSDB [In (%)].

Freatures	Methods	Acc			Abnormal						Normal
					Pre	Rec	Spe	F1			Pre	Rec	Spe	F1
	LR	83.46			80.01	75.28	88.47	77.57			85.38	88.47	75.28	86.90
	kNN	86.32			82.28	81.55	89.24	81.92			88.75	89.24	81.55	88.99
	RF	86.08			80.36	83.87	87.43	82.08			89.84	87.43	83.87	88.62
Color histogram	Linear SVM	46.28			39.48	77.62	27.06	52.34			66.36	27.06	77.62	38.45
	Non-linear SVM	62.00			Null	0.00	100.00	0.00			62.00	100.00	0.00	76.54
	ANN	81.20			78.14	70.14	87.98	73.93			82.78	87.98	70.14	85.30
	LR	71.28			66.89	48.39	85.32	56.15			72.95	85.32	48.39	78.65
	kNN	76.43			68.19	71.17	79.65	69.65			81.84	79.65	71.17	80.73
	RF	77.60			69.36	73.57	80.08	71.40			83.17	80.08	73.57	81.60
	Linear SVM	58.54			47.45	84.62	42.55	60.80			81.86	42.55	84.62	55.99
Luminance histogram	Non-linear SVM	62.00			Null	0.00	100.00	0.00			62.00	100.00	0.00	76.54
	ANN	71.18			62.97	58.65	78.86	60.73			75.68	78.86	58.65	77.23
HOG	LR	61.78			33.72	0.58	99.30	1.15			61.97	99.30	0.58	76.31
	kNN	62.02			50.04	39.56	75.79	44.18			67.17	75.79	39.56	71.22
	Naive Bayesian	54.83			44.29	73.06	43.66	55.15			72.56	43.66	73.06	54.52
	RF	60.55			48.15	49.62	67.25	48.87			68.53	67.25	49.62	67.88
	Linear SVM	50.91			39.90	57.60	46.81	47.14			64.30	46.81	57.60	54.18
	Non-linear SVM	62.00			Null	0.00	100.00	0.00			62.00	100.00	0.00	76.54
	ANN	62.35			54.37	5.77	97.03	10.43			62.69	97.03	5.77	76.17
LBP	LR	73.34			67.20	58.29	82.56	62.43			76.35	82.56	58.29	79.34
	kNN	70.27			64.05	49.64	82.92	55.93			72.87	82.92	49.64	77.57
	Naive Bayesian	57.39			46.41	78.43	44.49	58.31			77.09	44.49	78.43	56.42
	RF	70.13			60.88	59.90	76.41	60.39			75.66	76.41	59.90	76.03
	Linear SVM	46.21			29.70	30.40	55.89	30.05			56.71	55.89	30.40	56.30
	Non-linear SVM	62.00			Null	0.00	100.00	0.00			62.00	100.00	0.00	76.54
	ANN	71.19			64.72	53.19	82.23	58.39			74.13	82.23	53.19	77.97
GLCM	LR	67.54			58.21	51.65	77.27	54.74			72.28	77.27	51.65	74.69
	kNN	69.79			61.98	53.04	80.05	57.16			73.56	80.05	53.04	76.67
	Naive Bayesian	61.40			49.52	80.77	49.53	61.39			80.77	49.53	80.77	61.41
	RF	71.15			61.42	64.72	75.09	63.03			77.64	75.09	64.72	76.34
	Linear SVM	66.66			55.02	67.30	66.28	60.54			76.78	66.28	67.30	71.14
	Non-linear SVM	69.43			60.08	58.27	76.27	59.16			74.88	76.27	58.27	75.57
	ANN	68.10			58.45	55.56	75.79	56.97			73.56	75.79	55.56	74.66

The bold text in the table indicates the highest value of the classification result of different classifiers for the same feature.