5-way Fine-grained Image Classification on CUB (Few-Shot Evaluation)
79.61-Shot AccuracySum-min
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Sum-minBackbone=SF12*2022.04 | 79.6 | 90.48 | |
| Min-minBackbone=SF12*2022.04 | 78.51 | 89.73 | |
| Match-sumBackbone=SF12*2022.04 | 77.95 | 88.93 | |
| MixtFSLBackbone=ResNet182022.04 | 73.94 | 86.01 | |
| Neg-MarginBackbone=ResNet182022.04 | 72.66 | 89.4 | |
| Sum-minBackbone=SF4-642022.04 | 72.09 | 87.05 | |
| ProtoNet+Backbone=ResNet182022.04 | 71.88 | 86.64 | |
| MELRBackbone=Conv4-642022.04 | 70.26 | 85.01 | |
| Min-minBackbone=SF4-642022.04 | 70.15 | 84.94 | |
| FEATBackbone=Conv4-642022.04 | 68.87 | 82.9 | |
| MAML+Backbone=ResNet182022.04 | 68.42 | 83.47 | |
| RelationNet+Backbone=ResNet182022.04 | 67.59 | 82.75 | |
| Match-sumBackbone=SF4-642022.04 | 67.35 | 83.82 | |
| Baseline++Backbone=ResNet182022.04 | 67.02 | 83.58 | |
| ProtoNetBackbone=Conv4-642022.04 | 64.42 | 81.82 | |
| RelationNetBackbone=Conv4-642022.04 | 62.45 | 76.11 | |
| MatchingNetBackbone=Conv4-642022.04 | 61.16 | 72.86 | |
| Robust-20Backbone=ResNet182022.04 | 58.67 | 75.62 | |
| MAMLBackbone=Conv4-642022.04 | 55.92 | 72.09 |