Mammography Classification on VinDr
0.866ROC AUCAttend What Matters
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Attend What MattersModality=Image-Text2026.04 | 0.866 | — | 54.5 | 66.5 | 80.7 | 90.3 | |
| TReg-SwinTModality=Vision only2026.04 | 0.858 | — | 53 | 55.1 | 80.6 | 90.2 | |
| M-C-B5Modality=Image-Text2026.04 | 0.858 | — | 50.8 | 65.4 | 83.5 | 89.9 | |
| ViT-BModality=Vision only, Backbone=DINOv2 [1], Resolution=1024x1024, Classification Head=Linear layer2026.04 | 0.83 | — | 50 | 61.4 | 77 | 86.9 | |
| M_VinDrMode=tr, Normalization=Default, Variant=None2025.08 | 0.8 | 0.84 | — | — | — | — | |
| M_VinDr(BNFC)Mode=tr, Normalization=Default, Variant=None2025.08 | 0.8 | 0.83 | — | — | — | — | |
| M_HCTP+VinDr(BNFC)Mode=tr, Normalization=Default, Variant=None2025.08 | 0.79 | 0.82 | — | — | — | — | |
| M_HCTP+VinDr(DA_BNFC)Mode=tr, Normalization=Default, Variant=None2025.08 | 0.79 | 0.82 | — | — | — | — | |
| ViT-AModality=Vision only, Backbone=DINOv2 [1], Resolution=448x448, Classification Head=Linear layer2026.04 | 0.79 | — | 41.1 | 55 | 71.2 | 84.3 | |
| ViT-CModality=Vision only, Backbone=DINOv2 [1], Resolution=448x448, Classification Head=DeIT head2026.04 | 0.784 | — | 31.1 | 43.7 | 67.2 | 82.4 | |
| M_HCTP+VinDrMode=tr, Normalization=Default, Variant=None2025.08 | 0.78 | 0.82 | — | — | — | — | |
| MMBCDModality=Image-Text2026.04 | 0.771 | — | 27 | 50 | 66.2 | 82.8 | |
| M_HCTP+VinDr(DA)Mode=tr, Normalization=Default, Variant=None2025.08 | 0.77 | 0.82 | — | — | — | — | |
| M_HCTP(BNFC)Mode=tr, Normalization=Default, Variant=None2025.08 | 0.74 | 0.78 | — | — | — | — | |
| M_HCTP+VinDrMode=tt, Normalization=Default, Variant=**2025.08 | 0.72 | 0.76 | — | — | — | — | |
| M_HCTPMode=tr, Normalization=Default, Variant=None2025.08 | 0.7 | 0.73 | — | — | — | — | |
| M_HCTP+VinDrMode=tt, Normalization=Default, Variant=*2025.08 | 0.7 | 0.75 | — | — | — | — | |
| M_NYUMode=tt, Normalization=Default, Variant=**2025.08 | 0.69 | 0.74 | — | — | — | — | |
| M_NYUMode=tt, Normalization=Default, Variant=*2025.08 | 0.68 | 0.73 | — | — | — | — | |
| M_HCTPMode=tt, Normalization=Default, Variant=*2025.08 | 0.68 | 0.71 | — | — | — | — | |
| M_HCTPMode=tt, Normalization=Default, Variant=**2025.08 | 0.68 | 0.71 | — | — | — | — | |
| M_NYUMode=tt, Normalization=[0, 1], Variant=*2025.08 | 0.67 | 0.71 | — | — | — | — | |
| M_NYUMode=tt, Normalization=[0, 1], Variant=**2025.08 | 0.67 | 0.72 | — | — | — | — | |
| M_VinDrMode=tt, Normalization=Default, Variant=*2025.08 | 0.67 | 0.73 | — | — | — | — | |
| M_VinDrMode=tt, Normalization=Default, Variant=**2025.08 | 0.67 | 0.73 | — | — | — | — | |
| M_VinDr(BNFC)Mode=tr, Normalization=Default, Variant=**2025.08 | 0.66 | 0.7 | — | — | — | — | |
| M_VinDr(BNFC)Mode=tr, Normalization=Default, Variant=*2025.08 | 0.65 | 0.7 | — | — | — | — | |
| M_HCTP(BNFC)Mode=tr, Normalization=Default, Variant=*2025.08 | 0.64 | 0.66 | — | — | — | — | |
| M_HCTP(BNFC)Mode=tr, Normalization=Default, Variant=**2025.08 | 0.64 | 0.66 | — | — | — | — | |
| M_HCTP+VinDr(BNFC)Mode=tt, Normalization=Default, Variant=*2025.08 | 0.64 | 0.69 | — | — | — | — | |
| M_HCTP+VinDr(BNFC)Mode=tt, Normalization=Default, Variant=**2025.08 | 0.64 | 0.69 | — | — | — | — | |
| XFMambaModality=Vision only2026.04 | 0.636 | — | 18.3 | 25.2 | 51.5 | 64.6 | |
| MedVAEModality=Vision only2026.04 | 0.575 | — | 20.6 | 23.7 | 41.9 | 60.1 | |
| M_NYUMode=tr, Normalization=Default, Variant=None2025.08 | 0.55 | 0.6 | — | — | — | — | |
| M_NYUMode=tr, Normalization=[0, 1], Variant=None2025.08 | 0.5 | 0.53 | — | — | — | — |