Metric: MAE (lower is better)
| # | Model↕ | MAE▲ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | FOCUS | 0.036 | No | FOCUS: Towards Universal Foreground Segmentation | 2025-01-09 | Code |
| 2 | SAM2-UNet | 0.043 | No | SAM2-UNet: Segment Anything 2 Makes Strong Encod... | 2024-08-16 | Code |
| 3 | SelfReformer-Swin | 0.049 | No | SelfReformer: Self-Refined Network with Transfor... | 2022-05-23 | Code |
| 4 | SelfReformer | 0.05 | No | SelfReformer: Self-Refined Network with Transfor... | 2022-05-23 | Code |
| 5 | EVPv2 | 0.053 | No | Explicit Visual Prompting for Universal Foregrou... | 2023-05-29 | Code |
| 6 | EVPv1 | 0.054 | No | Explicit Visual Prompting for Low-Level Structur... | 2023-03-20 | Code |
| 7 | LDF(ours) | 0.059 | No | Label Decoupling Framework for Salient Object De... | 2020-08-25 | Code |
| 8 | RCSB | 0.059 | No | Recursive Contour Saliency Blending Network for ... | 2021-05-28 | Code |
| 9 | F3Net | 0.061 | No | F3Net: Fusion, Feedback and Focus for Salient Ob... | 2019-11-26 | Code |
| 10 | F3Net | 0.086 | No | U$^2$-Net: Going Deeper with Nested U-Structure ... | 2020-05-18 | Code |