A Statistical-Textural-Features Based Approach for Classification of Solid Drugs Using Surface Microscopic Images

Table 7

Accuracies for test datasets using top 15 features.

Datasets	Number of features	FR-algorithm	SVM			KNN			NB
			Acc	Sn	Sp	Acc	Sn	Sp	Acc	Sn	Sp
A1	15	CS	50.00	30	57.14	71.43	80	28.57	50.00	30	57.14
A2	15	CS	46.67	33.33	66.67	60.00	44.44	83.33	60.00	44.44	83.33
A3	15	CS	40.00	33	50	60.00	67	50	60.00	44	83.33
T1	15	CS	80.00	66.67	100	60.00	66.67	50	53.33	33.33	83.33
T2	15	CS	68.75	44.44	100	68.75	77.78	57.14	62.50	44.44	85.71
T3	15	CS	55.56	75	40	77.78	87.5	70	55.56	37.5	70
W1	15	CS	87.50	100	71.43	81.25	77.78	85.71	68.75	77.78	57.14
W2	15	CS	81.25	100	57.14	68.75	66.67	71.43	56.25	44.44	71.43
W3	15	CS	88.24	100	75	76.47	77.78	75	70.59	44.44	100
A1, A2, and A3	15	CS	79.31	14.29	100	75.86	42.86	86.36	68.97	14.29	86.36
T1, T2, and T3	15	CS	54.29	85.71	46.43	82.86	71.43	85.71	68.57	14.29	82.14
W1, W2, and W3	15	CS	91.18	85.71	92.59	61.76	42.86	66.67	61.76	28.57	70.37
All	15	CS	69.86	85.71	68.18	84.93	14.29	92.42	78.08	14.29	84.85
A1	15	GR	28.57	20	28.57	64.29	80	14.29	57.14	40	57.14
A2	15	GR	60.00	55.56	66.67	53.33	66.67	33.33	66.67	66.67	66.67
A3	15	GR	53.33	67	33.33	40.00	33	50	66.67	78	50
T1	15	GR	60.00	66.67	50	60.00	44.44	83.33	66.67	66.67	66.67
T2	15	GR	93.75	88.89	100	56.25	44.44	71.43	81.25	66.67	100
T3	15	GR	61.11	87.5	40	38.89	50	30	61.11	75	50
W1	15	GR	93.75	100	85.71	81.25	88.89	71.43	87.50	88.89	85.71
W2	15	GR	87.50	88.89	85.71	56.25	55.56	57.14	75.00	77.78	71.43
W3	15	GR	88.24	100	75	82.35	88.89	75	88.24	100	75
A1, A2, and A3	15	GR	63.33	28.57	77.27	65.52	57.14	68.18	58.62	14.29	72.73
T1, T2, and T3	15	GR	57.14	71.43	53.57	51.43	85.71	42.86	51.43	57.14	50
W1, W2, and W3	15	GR	76.47	42.86	85.19	85.29	57.14	92.59	76.47	14.29	92.59
All	15	GR	71.23	14.29	77.27	82.19	28.57	87.88	67.12	14.29	72.73