Table 1 False positive and false negative rates for each class.
From: Application of deep learning for fruit defect recognition in Psidium guajava L
Classes | H. | C.I. | Burn | Brown | C.G. | Pe.P | P.P. | P.M. | Le. | B.D.H | M.D. | Thrips | P.D. | Total | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Random training set | Total number | 202 | 591 | 3 | 80 | 786 | 585 | 80 | 78 | 87 | 33 | 12 | 234 | 86 | 2857 |
Number of True Positive (TP) | 202 | 590 | 3 | 80 | 786 | 585 | 80 | 78 | 87 | 33 | 12 | 233 | 86 | 2852 | |
Number of false negatives (FN) | 0 | 1 | 0 | 0 | 0 | 2 | 0 | 1 | 0 | 0 | 0 | 1 | 0 | 5 | |
False negative rate (%) | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | |
Number of false positives (FP) | 0 | 17 | 0 | 5 | 44 | 34 | 6 | 12 | 6 | 0 | 0 | 28 | 9 | 161 | |
False positive rate (%) | 0 | 3 | 0 | 6 | 5 | 5 | 7 | 13 | 6 | 0 | 0 | 11 | 9 | 5 | |
Random test set | Total number | 79 | 220 | 3 | 21 | 334 | 262 | 49 | 32 | 29 | 5 | 8 | 102 | 43 | 1187 |
Number of True Positive (TP) | 57 | 153 | 2 | 13 | 252 | 37 | 25 | 17 | 3 | 3 | 72 | 23 | 307 | 964 | |
Number of false negatives | 22 | 67 | 1 | 8 | 27 | 10 | 12 | 7 | 12 | 2 | 5 | 30 | 20 | 223 | |
False negative rate (%) | 22 | 23 | 25 | 32 | 28 | 4 | 20 | 18 | 29 | 29 | 38 | 23 | 32 | 16 | |
Number of false positives | 18 | 22 | 0 | 12 | 18 | 13 | 6 | 7 | 11 | 1 | 0 | 25 | 18 | 151 | |
False positive rate (%) | 19 | 9 | 0 | 36 | 5 | 5 | 11 | 18 | 28 | 17 | 0 | 20 | 30 | 11 | |
Overall | Total number | 281 | 811 | 6 | 101 | 1120 | 847 | 129 | 110 | 116 | 38 | 20 | 336 | 129 | 4044 |
Number of false negatives | 22 | 68 | 1 | 8 | 27 | 12 | 12 | 8 | 12 | 2 | 5 | 31 | 20 | 228 | |
False negative rate (%) | 7 | 8 | 14 | 7 | 2 | 1 | 9 | 7 | 9 | 5 | 20 | 8 | 13 | 5 | |
Number of false positives | 18 | 39 | 0 | 17 | 62 | 47 | 12 | 19 | 17 | 1 | 0 | 53 | 27 | 312 | |
False positive rate (%) | 6 | 5 | 0 | 14 | 5 | 5 | 9 | 15 | 13 | 3 | 0 | 14 | 17 | 7 | |
