[Retracted] Tobacco Leaves Disease Identification and Spot Segmentation Based on the Improved ORB Algorithm

Table 7

Experimental results of feature selection of particle swarm optimization for 15 times.

Number of experiments	Binary representation of feature-preferred combinations	Adaptability	Cross-validation rate (%)	Verification set recognition rate (%)
1	0101101100101101101001010101	94.23	88.57	94.00
2	1001011000010101001010001011	94.15	88.69	93.67
3	0001001010110110101001111010	98.65	89.23	91.00
4	1101010011100001001100100011	95.56	87.87	95.67
5	0001100110010111110101010110	92.15	92.66	8933
6	1000100100010000101011011101	93.26	86.71	92.67
7	0010100110110110111001101111	94.31	87.24	93.00
8	1001100100011001101011001111	94.11	89.31	94.00
9	0001100101001111001100111100	95.68	88.65	92.65
10	1000010001101010101100101011	95.26	93.31	96.00
11	1001010010001011011010111001	94.37	92.33	94.33
12	1101010001001101100100101010	92.31	89.23	93.00
13	1101010001001101100100101010	91.36	87.56	92.00
14	1101110101001100110010010011	95.15	86.26	89.23
15	0001010101010010011001010101	98.26	91.22	97.65