Multi-label classification on Open Images v4 (test)
35.3Precision (K=10)Semantic Diversity Learning
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Semantic Diversity LearningTask=GZSL, Backbone=TResNet-M, Pre-training=ImageNet, M (Number of principal directions)=72021.05 | 35.3 | 75.3 | 37.8 | 32.9 | 40.8 | 23.6 | 54.5 | |
| LESATask=GZSL, M (Number of principal directions)=102021.05 | 16.2 | 45.4 | 17.4 | 14.3 | 18.9 | 10.2 | 23.9 | |
| Fast0TagTask=GZSL2021.05 | 14.8 | 45.2 | 16 | 12.9 | 17.3 | 9.3 | 21.5 | |
| Semantic Diversity LearningTask=ZSL, Backbone=TResNet-M, Pre-training=ImageNet, M (Number of principal directions)=72021.05 | 6.1 | 62.9 | 10.7 | 8.3 | 47 | 4.4 | 68.1 | |
| LabelEMTask=GZSL2021.05 | 4.8 | 45.2 | 5.2 | 5.1 | 5.6 | 3.7 | 8.5 | |
| CONSETask=GZSL2021.05 | 2.4 | 43.5 | 2.6 | 2.4 | 2.8 | 1.7 | 3.9 | |
| LESATask=ZSL, M (Number of principal directions)=102021.05 | 0.7 | 41.7 | 1.4 | 1 | 25.6 | 0.5 | 37.4 | |
| Fast0TagTask=ZSL2021.05 | 0.3 | 41.2 | 0.7 | 0.6 | 12.6 | 0.3 | 21.3 | |
| CONSETask=ZSL2021.05 | 0.2 | 40.4 | 0.4 | 0.3 | 7.3 | 0.2 | 11.3 | |
| LabelEMTask=ZSL2021.05 | 0.2 | 40.5 | 0.5 | 0.4 | 8.7 | 0.2 | 15.8 |