Building Change Detection on LEVIR building change detection (test)
93.81PrecisionCEECNet V2
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| CEECNet V2depth (d)=5, loss strategy=evolving2020.09 | 93.81 | 89.92 | 91.83 | 91.42 | 84.89 | |
| CEECNet V1depth (d)=5, loss strategy=evolving2020.09 | 93.73 | 89.93 | 91.79 | 91.38 | 84.82 | |
| FracTAL ResNetdepth (d)=10, loss strategy=evolving2020.09 | 93.63 | 90.04 | 91.8 | 91.39 | 84.84 | |
| FracTAL ResNetdepth (d)=5, loss strategy=evolving2020.09 | 93.6 | 89.38 | 91.44 | 91.02 | 84.23 | |
| FracTAL ResNetdepth (d)=0, loss strategy=evolving2020.09 | 93.5 | 89.79 | 91.61 | 91.2 | 84.51 | |
| CEECNet V1depth (d)=5, loss strategy=static2020.09 | 93.36 | 89.46 | 91.37 | 90.94 | 84.1 | |
| Chen and Shi2020.09 | 83.8 | 91 | 87.3 | — | — |