Icon Classification on Icon645
62.93Total AccuracyResNet101 + LDAM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ResNet101 + LDAMBackbone=ResNet101, Loss Function=LDAM loss2021.10 | 62.93 | 70.29 | 70.5 | 47.51 | |
| ResNet32 + LDAMBackbone=ResNet32, Loss Function=LDAM loss2021.10 | 42.65 | 55.68 | 46.42 | 24.94 | |
| ResNet32 + Focal LossBackbone=ResNet32, Loss Function=Focal loss2021.10 | 32.8 | 51.59 | 36.51 | 8.94 | |
| ResNet32 + CBBackbone=ResNet32, Loss Function=CB loss2021.10 | 27.91 | 19.66 | 36.51 | 33.53 |