Text+object to image generation on MoCA
2.575NIQEMoGen
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| MoGenInput=Text + Object2026.01 | 2.575 | 0.349 | 93.01 | 4.801 | 4.448 | 0.762 | — | 68.91 | 83.98 | |
| MoGenInput=Text + Object + Bounding Box2026.01 | 2.593 | 0.351 | 92.95 | 4.785 | 4.514 | 0.723 | 0.675 | 76.26 | 91.52 | |
| Omnigen2Input=Text + Object2026.01 | 2.61 | 0.332 | 90.82 | 4.891 | 4.517 | 0.697 | — | 33.81 | 8.94 | |
| XverseInput=Text + Object2026.01 | 2.624 | 0.328 | 91.35 | 4.702 | 4.376 | 0.705 | — | 21.64 | 5.72 | |
| MS-diffusionInput=Text + Object2026.01 | 2.672 | 0.306 | 85.38 | 4.582 | 4.129 | 0.654 | — | 11.72 | 1.36 | |
| MS-diffusionInput=Text + Object + Bounding Box2026.01 | 2.797 | 0.336 | 87.05 | 4.531 | 4.214 | 0.671 | 0.503 | 26.29 | 7.16 | |
| Emu2Input=Text + Object + Bounding Box2026.01 | 2.916 | 0.327 | 85.73 | 4.479 | 4.192 | 0.662 | 0.458 | 23.81 | 1.32 |