Few-Shot Class-Incremental Learning on SAR Target Recognition Dataset
100Accuracy (t=0)NC-FSCIL
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| NC-FSCILPub. Year=ICLR 23'2026.06 | 100 | 88.91 | 89.83 | 77.31 | 75.86 | 72.04 | 74.16 | 75.15 | 24.85 | |
| Proposed-ResNetPub. Year=-, Backbone=ResNet2026.06 | 99.88 | 93.01 | 95.95 | 86.68 | 80.93 | 78.81 | 77.15 | 71.99 | 27.89 | |
| FACTPub. Year=CVPR 22'2026.06 | 99.71 | 99.57 | 91.66 | 80.38 | 80.99 | 79.79 | 77.14 | 75.49 | 24.22 | |
| ProposedPub. Year=-2026.06 | 99.69 | 96.15 | 87.45 | 80.64 | 84.58 | 82.27 | 78.14 | 76.46 | 23.23 | |
| WaRPPub. Year=ICLR 23'2026.06 | 99.45 | 98.4 | 90.66 | 84.63 | 76.78 | 76.32 | 75.06 | 71.62 | 27.83 | |
| PriViLegePub. Year=CVPR 24'2026.06 | 81.06 | 84.47 | 87.37 | 79.42 | 79.73 | 78.38 | 77.15 | 74.8 | 6.26 |