Table 3 Evaluation results of architectures with dual-encoder and single-encoder.
From: Urban change detection of remote sensing images via deep-feature extraction
Models | Dataset | Pre_c (%) | Rec_c (%) | Pre_nc (%) | Rec_nc (%) | F1 (%) | Kappa (%) |
|---|---|---|---|---|---|---|---|
Method with dual-encoder | OSCD CD_Data_GZ | 43.22 58.29 | 57.13 56.03 | 97.73 94.36 | 96.09 94.83 | 49.21 57.14 | 46.18 51.74 |
Method with single-encoder | OSCD CD_Data_GZ | 34.61 56.74 | 25.61 52.87 | 96.18 93.98 | 97.48 94.81 | 29.44 54.74 | 26.13 48.36 |
