Text-to-Image Generation on Pathological T2I/I2I Merged (test)
484.38FIDUniPath
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| UniPathUnif. Model=true, Fine-tuned=false, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 484.38 | 0.192 | 0.538 | 7.55 | 21.61 | 10.22 | 8.93 | |
| PixcellUnif. Model=false, Fine-tuned=false, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 929.3 | 0.259 | 0.524 | — | — | — | — | |
| PathLDMUnif. Model=false, Fine-tuned=false, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 1,126.36 | 0.376 | 0.483 | 0.15 | 0.73 | 0.18 | 0.3 | |
| Show-o2Unif. Model=true, Fine-tuned=true, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 1,398.52 | 0.415 | 0.545 | 8.41 | 22.77 | 10.93 | 9.42 | |
| UniMedVLUnif. Model=true, Fine-tuned=false, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 1,435.07 | 0.363 | 0.52 | 3.82 | 11.52 | 5.31 | 5.14 | |
| SD1.5Unif. Model=false, Fine-tuned=true, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 1,804.69 | 0.519 | 0.483 | 0.66 | 1.91 | 0.8 | 0.85 | |
| BLIP3oUnif. Model=true, Fine-tuned=true, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 2,008.75 | 0.55 | 0.455 | 1.55 | 5.87 | 2.39 | 2.58 | |
| SDXLUnif. Model=false, Fine-tuned=true, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 2,570.19 | 0.602 | 0.445 | 0.34 | 1.2 | 0.6 | 0.77 | |
| Pixart-αUnif. Model=false, Fine-tuned=true, Visual Fidelity Backbone=Virchow2, Text-Image Alignment Backbone=MUSK2025.12 | 2,574.85 | 0.685 | 0.482 | 1.14 | 4.54 | 1.58 | 1.74 |