Instance Segmentation on StomataSeg 1.0 (test)
46.1Overall APMask R-CNN
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Mask R-CNNBackbone=ConvNeXt-V2-Base, Preprocessing=Patches2026.01 | 46.1 | 87.6 | 33.1 | 82.7 | 48.9 | 89.9 | 56.3 | 90.1 | |
| Mask R-CNNBackbone=PVTv2-B2, Preprocessing=Patches2026.01 | 44.1 | 83 | 33.9 | 71.6 | 44.7 | 88.5 | 53.7 | 89 | |
| Mask R-CNNBackbone=Swin-T, Preprocessing=Patches2026.01 | 43.9 | 83.5 | 35.4 | 77.8 | 45 | 87.2 | 51.3 | 85.3 | |
| Mask R-CNNBackbone=ResNet-50, Preprocessing=Patches2026.01 | 43.8 | 84.7 | 32.2 | 78.2 | 45.2 | 86.8 | 54 | 89.1 | |
| HTCBackbone=ResNet-50, Preprocessing=Patches2026.01 | 42 | 79.7 | 32.3 | 76 | 42.7 | 81 | 51 | 82 | |
| Mask2FormerBackbone=ResNet-50, Preprocessing=Patches2026.01 | 35.8 | 74.6 | 22.4 | 61.2 | 38.8 | 81.5 | 46.2 | 81.1 | |
| Mask R-CNNBackbone=ConvNeXt-V2-Base, Preprocessing=Original images2026.01 | 28.3 | 50.7 | 0.59 | 0.99 | 33.91 | 67.96 | 50.52 | 83.12 | |
| Mask R-CNNBackbone=PVTv2-B2, Preprocessing=Original images2026.01 | 26.9 | 50.9 | 0 | 0 | 31.86 | 67.92 | 48.9 | 84.72 | |
| HTCBackbone=ResNet-50, Preprocessing=Original images2026.01 | 26.36 | 51.15 | 0 | 0 | 35.56 | 76.27 | 43.54 | 77.16 | |
| Mask R-CNNBackbone=ResNet-50, Preprocessing=Original images2026.01 | 26.17 | 50.61 | 0 | 0 | 31.84 | 68.6 | 46.67 | 83.23 | |
| Mask R-CNNBackbone=Swin-T, Preprocessing=Original images2026.01 | 26 | 51.1 | 0 | 0 | 27.42 | 66.03 | 50.63 | 87.41 | |
| Mask2FormerBackbone=ResNet-50, Preprocessing=Original images2026.01 | 25.37 | 46.36 | 0.29 | 0.99 | 29.47 | 60.46 | 46.36 | 77.63 |