Table 3 Number of epochs, training time, and classifications performed by different architectures of RCNNs for sweet potato roots regarding shape, damage caused by insects, and skin color. UFMG (2022).
From: Convolutional neural networks in the qualitative improvement of sweet potato roots
Var | Architecture | Epochs | Time | TP | FN | FP | TN |
|---|---|---|---|---|---|---|---|
Shape | VGG-16 | 70 | 0:29:58.5 | 517 | 127 | 132 | 382 |
Inception-v3 | 25 | 0:12:00.3 | 543 | 101 | 158 | 356 | |
ResNet-50 | 76 | 0:31:48.1 | 526 | 118 | 160 | 354 | |
InceptionResNetV2 | 17 | 0:17:59.3 | 630 | 14 | 20 | 494 | |
EfficientNetB3 | 100 | 1:05:30.8 | 567 | 77 | 138 | 376 | |
Damage caused by insects | VGG-16 | 95 | 0:36:46.8 | 97 | 46 | 299 | 715 |
Inception-v3 | 18 | 0:07:47.7 | 111 | 32 | 99 | 915 | |
ResNet-50 | 24 | 0:28:16.3 | 73 | 70 | 251 | 763 | |
InceptionResNetV2 | 11 | 0:06:27.7 | 140 | 3 | 38 | 976 | |
EfficientNetB3 | 100 | 0:45:22.6 | 87 | 56 | 262 | 752 | |
Skin color | VGG-16 | 80 | 0:34:56.5 | 464 | 96 | 70 | 528 |
Inception-v3 | 31 | 0:15:07.8 | 455 | 105 | 184 | 414 | |
ResNet-50 | 98 | 1:20:17.9 | 481 | 79 | 98 | 500 | |
InceptionResNetV2 | 32 | 0:34:53.3 | 540 | 20 | 2 | 596 | |
EfficientNetB3 | 98 | 1:06:43.3 | 505 | 55 | 54 | 544 |
