Localization on ImageNet-1k (val)
0.334EHRCoIBA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CoIBADepth=DeiT3-L2025.07 | 0.334 | — | — | |
| IBABackbone=ViT-B, Training Protocol=Finetuned [20]2025.07 | 0.31 | — | — | |
| CoIBATraining Strategy=ViT, Backbone=ViT-B2025.07 | 0.305 | — | — | |
| CoIBATraining Strategy=ViT, Backbone=ViT-L2025.07 | 0.304 | — | — | |
| IIADepth=DeiT3-L2025.07 | 0.304 | — | — | |
| CoIBATraining Strategy=ViT*, Backbone=ViT-B2025.07 | 0.3 | — | — | |
| GenericDepth=DeiT3-L2025.07 | 0.297 | — | — | |
| Chefer-LRPBackbone=ViT-B, Training Protocol=Finetuned [20]2025.07 | 0.297 | — | — | |
| IBABackbone=ViT-B, Training Protocol=Non-finetuned2025.07 | 0.297 | — | — | |
| IIABackbone=ViT-B, Training Protocol=Finetuned [20]2025.07 | 0.293 | — | — | |
| BeyondTraining Strategy=ViT*, Backbone=ViT-B2025.07 | 0.286 | — | — | |
| GenericBackbone=ViT-B, Training Protocol=Finetuned [20]2025.07 | 0.285 | — | — | |
| CoIBABackbone=DeiT-B, Training Protocol=Finetuned [20]2025.07 | 0.281 | — | — | |
| Chefer-LRPBackbone=ViT-B, Training Protocol=Non-finetuned2025.07 | 0.277 | — | — | |
| Chefer-LRPBackbone=DeiT-B, Training Protocol=Finetuned [20]2025.07 | 0.274 | — | — | |
| CoIBAPatch size=82025.07 | 0.272 | — | — | |
| IIABackbone=DeiT-B, Training Protocol=Finetuned [20]2025.07 | 0.271 | — | — | |
| CoIBABackbone=DeiT-B, Training Protocol=Non-finetuned2025.07 | 0.269 | — | — | |
| CoIBATraining Strategy=ViT*, Backbone=ViT-L2025.07 | 0.265 | — | — | |
| CoIBAPatch size=322025.07 | 0.263 | — | — | |
| BeyondBackbone=ViT-B, Training Protocol=Finetuned [20]2025.07 | 0.263 | — | — | |
| IBATraining Strategy=ViT, Backbone=ViT-B2025.07 | 0.258 | — | — | |
| IIABackbone=ViT-B, Training Protocol=Non-finetuned2025.07 | 0.256 | — | — | |
| CoIBABackbone=ViT-B, Training Protocol=Finetuned [20]2025.07 | 0.253 | — | — | |
| Chefer-LRPBackbone=DeiT-B, Training Protocol=Non-finetuned2025.07 | 0.252 | — | — | |
| BeyondBackbone=ViT-B, Training Protocol=Non-finetuned2025.07 | 0.252 | — | — | |
| IBABackbone=DeiT-B, Training Protocol=Finetuned [20]2025.07 | 0.25 | — | — | |
| CoIBABackbone=ViT-B, Training Protocol=Non-finetuned2025.07 | 0.248 | — | — | |
| Chefer-LRPTraining Strategy=ViT, Backbone=ViT-B2025.07 | 0.247 | — | — | |
| CoIBADepth=ViT-L2025.07 | 0.247 | — | — | |
| IBAPatch size=322025.07 | 0.246 | — | — | |
| IBADepth=DeiT3-L2025.07 | 0.245 | — | — | |
| BeyondPatch size=322025.07 | 0.243 | — | — | |
| ViT-CXBackbone=ViT-B, Training Protocol=Non-finetuned2025.07 | 0.24 | — | — | |
| IIABackbone=DeiT-B, Training Protocol=Non-finetuned2025.07 | 0.239 | — | — | |
| GenericTraining Strategy=ViT, Backbone=ViT-B2025.07 | 0.237 | — | — | |
| IIATraining Strategy=ViT, Backbone=ViT-B2025.07 | 0.237 | — | — | |
| Chefer-LRPDepth=ViT-L2025.07 | 0.236 | — | — | |
| ViT-CXPatch size=322025.07 | 0.236 | — | — | |
| IBATraining Strategy=ViT, Backbone=ViT-L2025.07 | 0.235 | — | — | |
| IBABackbone=DeiT-B, Training Protocol=Non-finetuned2025.07 | 0.235 | — | — | |
| GenericTraining Strategy=ViT, Backbone=ViT-L2025.07 | 0.234 | — | — | |
| IIATraining Strategy=ViT, Backbone=ViT-L2025.07 | 0.234 | — | — | |
| ViT-CXTraining Strategy=ViT, Backbone=ViT-L2025.07 | 0.234 | — | — | |
| ViT-CXTraining Strategy=ViT*, Backbone=ViT-B2025.07 | 0.232 | — | — | |
| GenericBackbone=DeiT-B, Training Protocol=Finetuned [20]2025.07 | 0.232 | — | — | |
| GenericBackbone=ViT-B, Training Protocol=Non-finetuned2025.07 | 0.229 | — | — | |
| ViT-CXBackbone=ViT-B, Training Protocol=Finetuned [20]2025.07 | 0.229 | — | — | |
| Chefer-LRPTraining Strategy=ViT, Backbone=ViT-L2025.07 | 0.226 | — | — | |
| IBATraining Strategy=ViT*, Backbone=ViT-B2025.07 | 0.226 | — | — | |
| BeyondTraining Strategy=ViT, Backbone=ViT-B2025.07 | 0.225 | — | — | |
| ViT-CXTraining Strategy=ViT*, Backbone=ViT-L2025.07 | 0.221 | — | — | |
| BeyondTraining Strategy=ViT*, Backbone=ViT-L2025.07 | 0.218 | — | — | |
| BeyondBackbone=DeiT-B, Training Protocol=Non-finetuned2025.07 | 0.216 | — | — | |
| IBADepth=ViT-L2025.07 | 0.213 | — | — | |
| ViT-CXTraining Strategy=ViT, Backbone=ViT-B2025.07 | 0.212 | — | — | |
| ViT-CXDepth=ViT-L2025.07 | 0.212 | — | — | |
| ViT-CXBackbone=DeiT-B, Training Protocol=Non-finetuned2025.07 | 0.212 | — | — | |
| IBAPatch size=82025.07 | 0.211 | — | — | |
| ViT-CXBackbone=DeiT-B, Training Protocol=Finetuned [20]2025.07 | 0.21 | — | — | |
| BeyondBackbone=DeiT-B, Training Protocol=Finetuned [20]2025.07 | 0.208 | — | — | |
| ViT-CXDepth=DeiT3-L2025.07 | 0.207 | — | — | |
| BeyondDepth=DeiT3-L2025.07 | 0.207 | — | — | |
| BeyondTraining Strategy=ViT, Backbone=ViT-L2025.07 | 0.205 | — | — | |
| GenericBackbone=DeiT-B, Training Protocol=Non-finetuned2025.07 | 0.203 | — | — | |
| BeyondPatch size=82025.07 | 0.202 | — | — | |
| IIAPatch size=82025.07 | 0.201 | — | — | |
| BeyondDepth=ViT-L2025.07 | 0.201 | — | — | |
| GenericTraining Strategy=ViT*, Backbone=ViT-B2025.07 | 0.2 | — | — | |
| IBATraining Strategy=ViT*, Backbone=ViT-L2025.07 | 0.191 | — | — | |
| ViT-CXPatch size=82025.07 | 0.187 | — | — | |
| IIADepth=ViT-L2025.07 | 0.176 | — | — | |
| GenericDepth=ViT-L2025.07 | 0.173 | — | — | |
| GenericPatch size=82025.07 | 0.169 | — | — | |
| IIAPatch size=322025.07 | 0.15 | — | — | |
| GenericPatch size=322025.07 | 0.143 | — | — | |
| IIATraining Strategy=ViT*, Backbone=ViT-L2025.07 | 0.11 | — | — | |
| IIATraining Strategy=ViT*, Backbone=ViT-B2025.07 | 0.107 | — | — | |
| GenericTraining Strategy=ViT*, Backbone=ViT-L2025.07 | 0.103 | — | — | |
| AttnLRPBackbone=ViT-B2026.02 | — | 58.4 | 60.75 | |
| AttnLRPBackbone=DeiT-B2026.02 | — | 54.63 | 68.16 | |
| AttnLRPBackbone=DeiT-III-B2026.02 | — | 53.84 | 77.65 | |
| AttnLRPBackbone=DINO-B2026.02 | — | 37.49 | 75.98 | |
| Chefer-LRPBackbone=ViT-B2026.02 | — | 54.98 | 80.82 | |
| Chefer-LRPBackbone=DeiT-B2026.02 | — | 55.47 | 79.62 | |
| Chefer-LRPBackbone=DeiT-III-B2026.02 | — | 52.27 | 81.94 | |
| Chefer-LRPBackbone=DINO-B2026.02 | — | 49.99 | 81.56 | |
| DAVEBackbone=ViT-B2026.02 | — | 60.19 | 78.6 | |
| DAVEBackbone=DeiT-B2026.02 | — | 63.52 | 82.23 | |
| DAVEBackbone=DeiT-III-B2026.02 | — | 65.76 | 82.43 | |
| DAVEBackbone=DINO-B2026.02 | — | 51.33 | 83.38 | |
| I×GBackbone=ViT-B2026.02 | — | 32.67 | 55.33 | |
| I×GBackbone=DeiT-B2026.02 | — | 30.25 | 62.72 | |
| I×GBackbone=DeiT-III-B2026.02 | — | 30.01 | 67.32 | |
| I×GBackbone=DINO-B2026.02 | — | 33.28 | 69.98 | |
| IntGradBackbone=ViT-B2026.02 | — | 39.86 | 58.51 | |
| IntGradBackbone=DeiT-B2026.02 | — | 36.11 | 64.22 | |
| IntGradBackbone=DeiT-III-B2026.02 | — | 31.68 | 68.12 | |
| IntGradBackbone=DINO-B2026.02 | — | 36.98 | 72.15 | |
| LeGradBackbone=ViT-B2026.02 | — | 47.71 | 80.06 |