Instruction-guided Image Editing on HumanEdit Counting (Quantitative Metrics)
66.8Human Avg ScoreHuman Avg
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Human Avg2026.02 | 66.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Background Consistency2026.02 | — | — | 0.554 | — | — | — | — | — | — | — | — | — | — | |
| Clip Image Norm2026.02 | — | — | — | 0.925 | — | — | — | — | — | — | — | — | — | |
| Clip Text Norm2026.02 | — | — | — | — | 0.619 | — | — | — | — | — | — | — | — | |
| Dino Image Norm2026.02 | — | — | — | — | — | 0.809 | — | — | — | — | — | — | — | |
| L1 Error2026.02 | — | — | — | — | — | — | 0.382 | — | — | — | — | — | — | |
| L2 Error2026.02 | — | — | — | — | — | — | — | 0.294 | — | — | — | — | — | |
| Lpips2026.02 | — | — | — | — | — | — | — | — | 0.526 | — | — | — | — | |
| Mask Lpips2026.02 | — | — | — | — | — | — | — | — | — | 0.526 | — | — | — | |
| Mask Ssim2026.02 | — | — | — | — | — | — | — | — | — | — | 55.4 | — | — | |
| MLLM-as-a-Judgeprompt=Main2026.02 | — | 0.731 | — | — | — | — | — | — | — | — | — | — | — | |
| Psnr2026.02 | — | — | — | — | — | — | — | — | — | — | — | 0.515 | — | |
| Ssim2026.02 | — | — | — | — | — | — | — | — | — | — | — | — | 55.4 |