Fake Image Detection on HPE-Bench Text2LIVE 1.0
94.74AccOurs
Evaluation results
| Method | Links | ||
|---|---|---|---|
| OursFull Method=Layer-Selective MLLM framework2026.01 | 94.74 | 90 | |
| AIDECategory=AI-Generated Image Detection Methods, Fine-tuned=true2026.01 | 83.75 | 91.89 | |
| UnivCategory=AI-Generated Image Detection Methods, Fine-tuned=true2026.01 | 73.75 | 87.32 | |
| FakeShieldCategory=Deepfake Detection Methods, Fine-tuned=true2026.01 | 72.5 | 84.06 | |
| LLaVA-NeXTCategory=Multimodal Large Language Models, Model Scale=8B2026.01 | 70.97 | 57.5 | |
| LLama3.2-VisionCategory=Multimodal Large Language Models, Model Scale=11B2026.01 | 67.5 | 84.06 | |
| Qwen3-VLCategory=Multimodal Large Language Models, Model Scale=8B2026.01 | 67.5 | 80.6 | |
| HifiNetCategory=Deepfake Detection Methods, Fine-tuned=true2026.01 | 66.25 | 82.35 | |
| LagradCategory=AI-Generated Image Detection Methods, Fine-tuned=true2026.01 | 65 | 78.79 | |
| CNNSpotCategory=AI-Generated Image Detection Methods, Fine-tuned=true2026.01 | 58.75 | 73.02 | |
| DeepSeekVL2Category=Multimodal Large Language Models, Model Scale=small2026.01 | 58.75 | 75 | |
| MVSSNetCategory=Deepfake Detection Methods, Fine-tuned=true2026.01 | 57.5 | 66.67 | |
| MiniCPM-V2.6Category=Multimodal Large Language Models, Model Scale=8B2026.01 | 55 | 66.67 | |
| LLaVA-1.6Category=Multimodal Large Language Models, Model Scale=7B2026.01 | 53.75 | 66.67 | |
| Random Choice2026.01 | 50 | 50 | |
| PSCCNetCategory=Deepfake Detection Methods, Fine-tuned=true2026.01 | 50 | 82.35 | |
| mPLUG-Owl3Category=Multimodal Large Language Models, Model Scale=7B2026.01 | 47.5 | 66.67 | |
| InternVL3.5Category=Multimodal Large Language Models, Model Scale=8B2026.01 | 46.25 | 54.55 | |
| Ovis2.5Category=Multimodal Large Language Models, Model Scale=9B2026.01 | 42.5 | 54.55 |