Image-level manipulation detection on DEFACTO 12k
78.3AUCDefakerOne
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DefakerOneModel Category=MLLMs2026.05 | 78.3 | — | — | — | |
| TruforModel Category=Vision Models2026.05 | 64.3 | — | — | — | |
| MesorchModel Category=Vision Models2026.05 | 63.9 | — | — | — | |
| MVSS-NetDecision threshold=0.52021.04 | 57.3 | 81.7 | 26.8 | 40.4 | |
| MVSS-NetDecision threshold=0.52021.12 | 57.3 | 81.7 | 26.8 | 40.4 | |
| CR-CNNDecision threshold=0.52021.04 | 56.7 | 77.4 | 26.7 | 39.7 | |
| CR-CNNDecision threshold=0.52021.12 | 56.7 | 77.4 | 26.7 | 39.7 | |
| DRCTModel Category=Vision Models2026.05 | 55.3 | — | — | — | |
| FCNDecision threshold=0.52021.04 | 55.1 | 71.1 | 33.8 | 45.8 | |
| FCNDecision threshold=0.52021.12 | 55.1 | 71.1 | 33.8 | 45.8 | |
| DTDModel Category=Vision Models2026.05 | 55.1 | — | — | — | |
| ManrTra-NetDecision threshold=0.52021.04 | 54.3 | 100 | 0 | 0 | |
| ManrTra-NetDecision threshold=0.52021.12 | 54.3 | 100 | 0 | 0 | |
| CAT-NetDecision threshold=0.52021.12 | 54.3 | 34.2 | 72.5 | 46.5 | |
| CDFAModel Category=Vision Models2026.05 | 53.6 | — | — | — | |
| EffortModel Category=Vision Models2026.05 | 53.5 | — | — | — | |
| MVSS-Net++Decision threshold=0.52021.12 | 53.1 | 37.3 | 66.6 | 47.8 | |
| SPANDecision threshold=0.52021.12 | 50 | 100 | 0 | 0 | |
| FakeVLMModel Category=MLLMs2026.05 | 50 | — | — | — | |
| H-LSTMDecision threshold=0.52021.12 | 49.9 | 99.9 | 0.1 | 0.2 | |
| GSR-NetDecision threshold=0.52021.04 | 45.6 | 91.4 | 0.1 | 0.2 | |
| GSR-NetDecision threshold=0.52021.12 | 45.6 | 91.4 | 0.1 | 0.2 | |
| FFDNModel Category=Vision Models2026.05 | 37.8 | — | — | — | |
| Ivy-xDetectorModel Category=MLLMs2026.05 | 30.9 | — | — | — | |
| ForensicMOEModel Category=MLLMs2026.05 | 1.7 | — | — | — | |
| VeritasModel Category=MLLMs2026.05 | 0.9 | — | — | — |