Hierarchical Relationship Prediction on ImageNet Instruments Novel Datasets
78.84Avg F1 ScoreTASK2BOX
Evaluation Results
| Method | Links | |
|---|---|---|
| TASK2BOXBox Dimension=5D, Input Representation=[μ, σ²] CLIP2024.03 | 78.84 | |
| TASK2BOXBox Dimension=5D, Input Representation=CLIP2024.03 | 76.95 | |
| TASK2BOXBox Dimension=3D, Input Representation=CLIP2024.03 | 73.06 | |
| TASK2BOXBox Dimension=3D, Input Representation=[μ, σ²] CLIP2024.03 | 66.48 | |
| MLP ClassifierInput Representation=[μ, σ²] CLIP2024.03 | 64.56 | |
| MLP ClassifierInput Representation=CLIP2024.03 | 62.24 | |
| TASK2BOXBox Dimension=2D, Input Representation=[μ, σ²] CLIP2024.03 | 48.35 | |
| TASK2BOXBox Dimension=2D, Input Representation=CLIP2024.03 | 47.76 | |
| MLP ClassifierInput Representation=FIM2024.03 | 41.22 | |
| TASK2BOXBox Dimension=5D, Input Representation=FIM2024.03 | 37.39 | |
| TASK2BOXBox Dimension=3D, Input Representation=FIM2024.03 | 24.7 | |
| Linear ClassifierInput Representation=FIM2024.03 | 22.98 | |
| Linear ClassifierInput Representation=CLIP2024.03 | 11.89 | |
| Linear ClassifierInput Representation=[μ, σ²] CLIP2024.03 | 11.87 | |
| KL DivergenceInput Representation=[μ, σ²] CLIP2024.03 | 10.41 | |
| TASK2BOXBox Dimension=2D, Input Representation=FIM2024.03 | 9.46 | |
| RandomInput Representation=FIM2024.03 | 5.02 | |
| Asymmetric EuclideanInput Representation=FIM2024.03 | 4.73 | |
| Asymmetric CosineInput Representation=CLIP2024.03 | 4.52 | |
| Asymmetric CosineInput Representation=[μ, σ²] CLIP2024.03 | 4.52 | |
| Asymmetric EuclideanInput Representation=CLIP2024.03 | 4.52 | |
| Asymmetric EuclideanInput Representation=[μ, σ²] CLIP2024.03 | 4.52 | |
| KL DivergenceInput Representation=FIM2024.03 | 0 | |
| Asymmetric CosineInput Representation=FIM2024.03 | 0 |