Table 5. Performance evaluation of HierbaNetV1 against pre-trained models on SorghumWeedDataset_Classification using accuracy, precision, recall, F1-score, and loss.
Model | Accuracya | Precision | Recall | F1-score | Loss |
---|---|---|---|---|---|
HierbaNetV1 (Proposed model) | 0.9861 | 0.9860 | 0.9862 | 0.9860 | 0.0700 |
Pre-trained models | |||||
InceptionV3 | 0.9791 | 0.9795 | 0.9795 | 0.9792 | 1.5472 |
VGG19 | 0.9698 | 0.9704 | 0.9702 | 0.9698 | 0.1444 |
ResNet152V2 | 0.9675 | 0.9685 | 0.9681 | 0.9676 | 1.4649 |
DenseNet201 | 0.9582 | 0.9601 | 0.9590 | 0.9583 | 1.0096 |
MobileNetV2 | 0.9327 | 0.9336 | 0.9321 | 0.9323 | 2.2053 |
SOTA architectures | |||||
Hybrid CNN-Transformer | 0.9606 | 0.9604 | 0.9605 | 0.9605 | 1.1503 |
DarkNet53 | 0.9397 | 0.9395 | 0.9397 | 0.9395 | 2.5037 |
RCNN | 0.9142 | 0.9140 | 0.9142 | 0.9140 | 3.0356 |
Note:
The table is organized based on accuracy within each section.