Table 11 Comparison between the proposed models and the state-of-the-art DL for flower detection.
From: Deep learning based approach for actinidia flower detection and gender assessment
Application | DL models | Results | Article | ||||
|---|---|---|---|---|---|---|---|
Precision (%) | Recall (%) | F1 score (%) | AP (%) | mAP (%) | |||
Flower | Flower | Flower | |||||
Flower and stamen detection | Modified YOLOv5s | 96.7 | 89.1 | – | – | 90.1 | |
Faster R-CNN + ResNet50 | 57.4 | 98.9 | – | – | 92.6 | ||
Faster R-CNN + VGG | 68.5 | 98.9 | – | – | 92.6 | ||
SDD + VGG | 76.6 | 87.4 | – | – | 82.3 | ||
SDD MobileNetv2 | 86.7 | 70.2 | – | – | 81.1 | ||
Flower and bud detection | YOLOv4 | – | – | – | 92.47 | 91.49 | |
YOLOV3 | – | – | – | 85.73 | 80.98 | ||
Flower detection | Faster R-CNN NAS | 96.8 | 68.0 | 79.0 | – | – | |
Faster R-CNN Inception v2 | 90.4 | 75.8 | 82.0 | – | – | ||
SDD Inception v2 | 78.5 | 61.2 | 68.1 | – | – | ||
Flower detection | Faster R-CNN Inception v2 | 91 | 80 | 85 | – | – | |
Flower and bud detection | YOLOv5l | – | – | – | 93.12 | 93.23 | |
Flower and gender detection | YOLOv5 | 95 | 94 | 94 | – | 86 | Proposed |
YOLOv8 | 96 | 93 | 94 | – | 85 | ||
DETR | 89 | 97 | 93 | – | 94 | ||
RT-DETR | 95 | 96 | 95 | – | 89 | ||
