Building Change Detection on WHU building change detection dataset
95.571PrecisionCEECNet V1
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| CEECNet V1FT similarity metric depth (d)=5, Loss strategy=evo, Input size=256 x 2562020.09 | 95.571 | 92.043 | 93.774 | 93.616 | 88.23 | |
| FracTAL ResNetFT similarity metric depth (d)=5, Loss strategy=evo, Input size=256 x 2562020.09 | 95.35 | 90.873 | 93.058 | 92.892 | 87.02 | |
| Cao et al.Input size=512 x 5122020.09 | 94 | 79.37 | 86.07 | — | — | |
| MS-FCN (M2)Input size=512 x 5122020.09 | 93.8 | 87.8 | 90.7 | — | 83 | |
| Mask-RCNN (M1)Input size=512 x 5122020.09 | 93.1 | 89.2 | 91.108 | — | 83.7 | |
| Liu et al.Input size=256 x 2562020.09 | 90.15 | 89.35 | 89.75 | — | 81.4 | |
| Chen et al.Input size=512 x 5122020.09 | 89.2 | 90.5 | 89.8 | — | — |