Table 4 Performance comparison of different models on various datasets.
From: The analysis of landscape design and plant selection under deep learning
Dataset | Model architecture | Accuracy of landscape element classification (%) | Accuracy of fine-grained PLIR (%) |
|---|---|---|---|
Sift flow | FCN-8s | 85.2 | 81.7 |
Improved FCN model | 89.7 | 86.5 | |
FCN + Faster R-CNN | 90.4 | 88.7 | |
ADE20K | FCN-8s | 78.3 | 75.1 |
Improved FCN model | 82.5 | 78.6 | |
FCN + Faster R-CNN | 84.7 | 81.2 | |
Cityscapes | FCN-8s | 79.9 | 77.2 |
Improved FCN model | 84.2 | 80.3 | |
FCN + Faster R-CNN | 86.1 | 82.5 |
