Pixel-wise segmentation on SEM images dataset of Cr-superalloys (test)
98.3AccuracyDT-SegNet
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DT-SegNet2026.06 | 98.3 | 87.8 | 92.8 | 81.9 | 94 | 90 | |
| SegFormervariant=B02026.06 | 98.1 | 87.6 | 88.9 | 78.3 | 93.3 | 87.7 | |
| SegFormervariant=B12026.06 | 97.7 | 84.1 | 91.9 | 77.1 | 92.4 | 87 | |
| DeepLabV3+2026.06 | 97.6 | 85.6 | 87.1 | 75.5 | 92.4 | 85.9 | |
| U-Net2026.06 | 96.4 | 80.5 | 73.8 | 63.9 | 92 | 71.3 | |
| UNet 3+2026.06 | 96.3 | 87.6 | 70.4 | 61.9 | 91.5 | 70 | |
| ilastikvariant=RF2026.06 | 93.9 | 63.7 | 72.7 | 49.5 | 82.4 | 65.2 | |
| Wekavariant=FRF2026.06 | 93.5 | 76.1 | 69.6 | 51.6 | 82.7 | 68 | |
| Wekavariant=MLP2026.06 | 92 | 58.7 | 79.5 | 48.6 | 70.9 | 64 | |
| ilastikvariant=LDA2026.06 | 86.8 | 40 | 82.7 | 35.9 | 65.7 | 51.6 | |
| ilastikvariant=SVC2026.06 | 75.5 | 49.4 | 56.2 | 23.5 | 68.8 | 35.3 |