BI-RADS-based malignancy classification on VinDr (test)
92AUCFPN-AbMIL
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| FPN-AbMIL# Trainable Params (M)=1.76, Level=Image2026.03 | 92 | 84.4 | 72 | |
| SIL-L GMIC-ResNet18# Trainable Params (M)=22.49, Level=Image2026.03 | 91.1 | 85.5 | 69.8 | |
| FPN-AbMIL (mean)# Trainable Params (M)=1.76, Level=Breast2026.03 | 91.1 | 83.1 | 64.3 | |
| MIL-PF# Trainable Params (M)=0.04, Level=Breast, Backbone=MedSigLIP, Aggregation=attn2026.03 | 91.1 | 85 | 68.3 | |
| FPN-SetTrans# Trainable Params (M)=5.38, Level=Image2026.03 | 90.8 | 82.7 | 67 | |
| MIL-PF# Trainable Params (M)=0.04, Level=Breast, Backbone=MedSigLIP, Aggregation=max2026.03 | 90.1 | 83.8 | 61.6 | |
| FPN-SetTrans (mean)# Trainable Params (M)=1.76, Level=Breast2026.03 | 90 | 81.1 | 55.9 | |
| GMIC# Trainable Params (M)=14.11, Level=Image2026.03 | 89.9 | 82.6 | 69 | |
| FPN-AbMIL (max)# Trainable Params (M)=5.38, Level=Breast2026.03 | 89.9 | 83 | 51.2 | |
| MIL-PF# Trainable Params (M)=0.05, Level=Breast, Backbone=DINOv2, Aggregation=attn2026.03 | 89.4 | 84.5 | 79.2 | |
| FPN-SetTrans (max)# Trainable Params (M)=5.38, Level=Breast2026.03 | 88.7 | 77.6 | 59.5 | |
| MIL-PF# Trainable Params (M)=0.05, Level=Breast, Backbone=DINOv2, Aggregation=max2026.03 | 88.2 | 81.1 | 71.2 | |
| ES-Attside# Trainable Params (M)=22.89, Level=Case2026.03 | 78.6 | 74.9 | 19.3 |