Image Attribution
Image attribution algorithms aim to identify important regions that are highly relevant to model decisions.
Papers
Showing 1–10 of 26 papers
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | SMDL-Attribution (ICLR version) | Insertion AUC score (ArcFace ResNet-101) | 0.67 | — | Unverified |
| 2 | HSIC-Attribution | Insertion AUC score (ArcFace ResNet-101) | 0.67 | — | Unverified |
| 3 | RISE | Insertion AUC score (ArcFace ResNet-101) | 0.65 | — | Unverified |
| 4 | LIME | Insertion AUC score (ArcFace ResNet-101) | 0.62 | — | Unverified |
| 5 | Kernel SHAP | Insertion AUC score (ArcFace ResNet-101) | 0.61 | — | Unverified |
| 6 | Saliency | Insertion AUC score (ArcFace ResNet-101) | 0.56 | — | Unverified |
| 7 | Integrated Gradients | Insertion AUC score (ArcFace ResNet-101) | 0.54 | — | Unverified |
| 8 | Grad-CAM | Insertion AUC score (ArcFace ResNet-101) | 0.47 | — | Unverified |