Image Classification, OOD, and Adversarial Detection on Oxford Flowers/Deep Weeds (test)
100ID Accuracy (%)C-EDL
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| C-EDLUncertainty Strategy=Monte Carlo2025.06 | 100 | 42.33 | 1.02 | 1.14 | |
| R-EDL2025.06 | 99.78 | 64.78 | 2.34 | 1.25 | |
| EDL++Uncertainty Strategy=Monte Carlo2025.06 | 99.72 | 54.2 | 6 | 5.55 | |
| C-EDLUncertainty Strategy=Metamorphic Transformation2025.06 | 99.59 | 47.5 | 1.78 | 1.93 | |
| EDL++Uncertainty Strategy=Metamorphic Transformation2025.06 | 99.2 | 51.73 | 5.61 | 6.94 | |
| S-EDL2025.06 | 99.09 | 61.7 | 8.13 | 11.56 | |
| EDL2025.06 | 99 | 63.67 | 8.87 | 13.57 | |
| H-EDL2025.06 | 98.97 | 63.17 | 6.01 | 8.61 | |
| I-EDL2025.06 | 98.89 | 60.73 | 8.17 | 12.74 | |
| Posterior Network2025.06 | 98.59 | 69.86 | 5.27 | 8.24 | |
| DA-EDL2025.06 | 92.38 | 71.4 | 3.76 | 4.69 |