Object Hallucination Detection on POPE-AOKVQA
58.8AccuracyVideo-LLaVA + VisAlign + VCD
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Video-LLaVA + VisAlign + VCDModel=Video-LLaVA + VisAlign + VCD2025.11 | 58.8 | 55.03 | 96.33 | 70.04 | |
| Video-LLaVA + VisAlignModel=Video-LLaVA + VisAlign2025.11 | 57.09 | 53.9 | 98.33 | 69.63 | |
| Video-LLaVA + VCDModel=Video-LLaVA + VCD2025.11 | 54.5 | 52.38 | 99.39 | 68.6 | |
| Video-LLaVAModel=Video-LLaVA2025.11 | 54.1 | 52.14 | 99.6 | 68.45 |