Image Editing on Image Editing Dataset (Counting)
4.4Unchanged RegionsMLLM-as-a-Judge
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MLLM-as-a-JudgeJudge=Our Judge2026.02 | 4.4 | 4.8 | 6.4 | 6.4 | 4.8 | 5.8 | 6.2 | 6.2 | 5.6 | 5.2 | 5.2 | 6.8 | 5.65 | |
| HumanJudge=Human2026.02 | 3.393 | 5.243 | 4.227 | 5.51 | 4.65 | 5.157 | 5.247 | 5.403 | 5.357 | 3.437 | 3.537 | 4.917 | 4.673 |