Image-level recognizability evaluation on IJB-C Image-level 18
0.6306SCTransFIRA (CCAS)
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| TransFIRA (CCAS)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.6306 | 0.0153 | 0.0285 | 0.123 | |
| TransFIRA (CCAS)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.6245 | 0.0144 | 0.0442 | 0.6659 | |
| TransFIRA (CCS)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.6245 | 0.0153 | 0.0285 | 0.1228 | |
| TransFIRA (CCS)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.6111 | 0.0141 | 0.0433 | 0.6659 | |
| MagFaceBackbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.603 | 0.0156 | 0.029 | 0.1237 | |
| CR-FIQA (L)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.5902 | 0.0156 | 0.029 | 0.1237 | |
| eDifFIQA (L)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.5803 | 0.0153 | 0.0286 | 0.1229 | |
| eDifFIQA (M)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.5715 | 0.0158 | 0.0292 | 0.1248 | |
| MagFaceBackbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.5684 | 0.0144 | 0.0439 | 0.6663 | |
| SDD-FIQABackbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.5682 | 0.0156 | 0.0291 | 0.1242 | |
| eDifFIQA (S)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.5663 | 0.0156 | 0.029 | 0.1236 | |
| CLIB-FIQABackbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.558 | 0.0154 | 0.0287 | 0.1231 | |
| CR-FIQA (L)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.5571 | 0.0144 | 0.044 | 0.667 | |
| SDD-FIQABackbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.5374 | 0.0144 | 0.0441 | 0.6668 | |
| eDifFIQA (L)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.5344 | 0.0141 | 0.0433 | 0.6674 | |
| eDifFIQA (M)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.5277 | 0.0146 | 0.0444 | 0.6683 | |
| eDifFIQA (S)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.5236 | 0.0144 | 0.0439 | 0.6672 | |
| DifFIQA (R)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.5193 | 0.0156 | 0.029 | 0.124 | |
| CLIB-FIQABackbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.5112 | 0.0142 | 0.0435 | 0.6677 | |
| DifFIQA (R)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.4729 | 0.0144 | 0.044 | 0.6688 | |
| CR-FIQA (S)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.4589 | 0.0154 | 0.0287 | 0.1232 | |
| CR-FIQA (S)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.4218 | 0.0142 | 0.0436 | 0.67 | |
| FaceQNetBackbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.3677 | 0.0154 | 0.0288 | 0.1239 | |
| FaceQNetBackbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.3607 | 0.0143 | 0.0438 | 0.6716 | |
| GraFIQs (L)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.344 | 0.0154 | 0.0287 | 0.1232 | |
| GraFIQs (L)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.3307 | 0.0142 | 0.0436 | 0.6733 | |
| GraFIQs (S)Backbone Architecture=CosFace, Backbone Pre-training Dataset=BRIAR2025.10 | 0.105 | 0.0143 | 0.0438 | 0.6793 | |
| GraFIQs (S)Backbone Architecture=ArcFace, Backbone Pre-training Dataset=WebFace2025.10 | 0.0911 | 0.0154 | 0.0288 | 0.1237 |