Image Classification on BreastMNIST v2 (test)
0.919AUCGoogle AutoML Vision
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Google AutoML Vision2026.02 | 0.919 | — | — | — | 86.1 | |
| ResNet-18Input Resolution=282026.02 | 0.901 | — | — | — | 86.3 | |
| ResNet-18Input Resolution=2242026.02 | 0.891 | — | — | — | 83.3 | |
| AutoKeras2026.02 | 0.871 | — | — | — | 83.1 | |
| ResNet-50Input Resolution=2242026.02 | 0.866 | — | — | — | 84.2 | |
| ResNet-50Input Resolution=282026.02 | 0.857 | — | — | — | 81.2 | |
| auto-sklearn2026.02 | 0.836 | — | — | — | 80.3 | |
| Topo-Med (MLP)2026.02 | 0.821 | — | — | — | 73.7 | |
| G-LoG bi-filtration (MLP)Sigma (σ)=02026.02 | 0.82 | — | — | — | 73.1 | |
| G-LoG bi-filtration (MLP)Sigma (σ)=0.52026.02 | 0.784 | — | — | — | 81.4 | |
| G-LoG bi-filtration (MLP)Sigma (σ)=12026.02 | 0.752 | — | — | — | 75 | |
| G-LoG bi-filtration (MLP)Sigma (σ)=1.52026.02 | 0.73 | — | — | — | 76.3 | |
| G*Selection strategy=G* (Proposed Ensemble Selection), Backbone=ResNet182025.03 | — | 0.755 | -7.6 | 32 | — | |
| G* (Oracle)Selection strategy=G* Oracle Selection, Backbone=ResNet182025.03 | — | 0.755 | -7.6 | 38 | — | |
| Naive (A)Selection strategy=Naive All Ensemble, Backbone=ResNet182025.03 | — | 0.714 | -12.6 | 110 | — | |
| Naive (R)Selection strategy=Naive Random Ensemble, Backbone=ResNet182025.03 | — | 0.664 | -18.7 | 32 | — | |
| RealTraining Data=Real, Backbone=ResNet182025.03 | — | 0.817 | — | — | — | |
| Single (average)Selection strategy=average of single GANs, Backbone=ResNet182025.03 | — | 0.407 | -50.2 | 1 | — | |
| Single (top-1)Selection strategy=top-1 ranking, Ranking source=Real test set performance, Backbone=ResNet182025.03 | — | 0.707 | -13.5 | 1 | — | |
| Single (top-5)Selection strategy=top-5 ranking, Ranking source=Real test set performance, Backbone=ResNet182025.03 | — | 0.697 | -14.7 | 1 | — |