Few-shot image classification on TieredImageNet ILSVRC-2012 (test)
86.5AccuracyProtoNet+LLS
Evaluation Results
| Method | Links | |
|---|---|---|
| ProtoNet+LLSBackbone=ResNet-12, Shots=5-shot2021.11 | 86.5 | |
| DeepEMDBackbone=ResNet-12, Shots=5-shot2021.11 | 86.03 | |
| Simple-ShotBackbone=WRN-28-10, Shots=5-shot2021.11 | 85.5 | |
| MCT (Inst)Backbone=ResNet-12, Shots=5-shot2021.11 | 85.29 | |
| FEATBackbone=ResNet-12, Shots=5-shot2021.11 | 84.79 | |
| CANBackbone=ResNet-12, Shots=5-shot2021.11 | 84.23 | |
| ProtoNet* (baseline)Backbone=ResNet-12, Shots=5-shot2021.11 | 83.67 | |
| MetaOptNet-SVMBackbone=ResNet-12, Shots=5-shot2021.11 | 81.56 | |
| ProtoNet+LLSBackbone=ResNet-12, Shots=1-shot2021.11 | 72.27 | |
| DeepEMDBackbone=ResNet-12, Shots=1-shot2021.11 | 71.16 | |
| FEATBackbone=ResNet-12, Shots=1-shot2021.11 | 70.8 | |
| CANBackbone=ResNet-12, Shots=1-shot2021.11 | 69.89 | |
| Simple-ShotBackbone=WRN-28-10, Shots=1-shot2021.11 | 69.75 | |
| MCT (Inst)Backbone=ResNet-12, Shots=1-shot2021.11 | 69.66 | |
| ProtoNet* (baseline)Backbone=ResNet-12, Shots=1-shot2021.11 | 67.67 | |
| MetaOptNet-SVMBackbone=ResNet-12, Shots=1-shot2021.11 | 65.99 |