Binary Glaucoma Classification on Fundus image dataset (5-fold stratified cross-validation)
99.42AccuracyGlaKG
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GlaKGInput Modality=KG-fusion, alpha=0.52026.07 | 99.42 | 100 | 99.06 | 99.53 | 99.88 | |
| GATInput Modality=KG subgraphs, Node features=64-dimensional2026.07 | 97.1 | 95.55 | 100 | 97.72 | 99.12 | |
| GCNInput Modality=KG subgraphs, Node features=64-dimensional2026.07 | 90.72 | 92.77 | 92.27 | 92.47 | 96.41 | |
| GBInput Modality=Image-only, Backbone=ResNet50, Estimators=1002026.07 | 88.69 | 87.71 | 95.09 | 91.24 | 94.87 | |
| LRInput Modality=Image-only, Backbone=ResNet502026.07 | 88.54 | 89.45 | 92.5 | 90.94 | 94.17 | |
| MLPInput Modality=Image-only, Backbone=ResNet50, Hidden Layers=(256, 64)2026.07 | 87.67 | 88.11 | 92.74 | 90.29 | 94.17 | |
| RFInput Modality=Image-only, Backbone=ResNet50, Estimators=2002026.07 | 87.23 | 83.88 | 98.36 | 90.53 | 94.23 |