Object State Classification on EPIC-STATES Novel Objects (test)
81.8mAPTSC+OHC
Evaluation results
| Method | Links | |
|---|---|---|
| TSC+OHCLinear classifier training data=100%2021.12 | 81.8 | |
| TSCLinear classifier training data=100%2021.12 | 80.2 | |
| TSC+OHCLinear classifier training data=12.5%2021.12 | 79.7 | |
| SimCLRLinear classifier training data=100%2021.12 | 77.1 | |
| Semantic supervision via EPIC action classificationLinear classifier training data=100%2021.12 | 77 | |
| ImageNet Pre-trainedLinear classifier training data=100%2021.12 | 74.5 | |
| TSCLinear classifier training data=12.5%2021.12 | 74.5 | |
| Semantic supervision via MIT States datasetLinear classifier training data=100%2021.12 | 73.9 | |
| SimCLRLinear classifier training data=12.5%2021.12 | 71.9 | |
| Semantic supervision via EPIC action classificationLinear classifier training data=12.5%2021.12 | 70.9 | |
| ImageNet Pre-trainedLinear classifier training data=12.5%2021.12 | 70.2 | |
| Semantic supervision via MIT States datasetLinear classifier training data=12.5%2021.12 | 70.1 | |
| SimCLR + TCNLinear classifier training data=100%2021.12 | 68.4 | |
| TCNLinear classifier training data=100%2021.12 | 63.9 | |
| SimCLR + TCNLinear classifier training data=12.5%2021.12 | 63.7 | |
| TCNLinear classifier training data=12.5%2021.12 | 56.1 |