Intent Prediction on Multimodal domestic cat dataset (LOVO)
77.72AccuracyPrior-Guided PoE (Full)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Prior-Guided PoE (Full)Evaluation Protocol=Leave-One-Video-Out (LOVO), alpha=1.02026.04 | 77.72 | 33.24 | 74.6 | |
| PoE-Ctx+AudEvaluation Protocol=Leave-One-Video-Out (LOVO)2026.04 | 75.31 | 35.77 | 75.49 | |
| LateFusion-WeightedEvaluation Protocol=Leave-One-Video-Out (LOVO)2026.04 | 73.76 | 32.02 | 74.54 | |
| LateFusion-AvgEvaluation Protocol=Leave-One-Video-Out (LOVO)2026.04 | 73.69 | 33.88 | 76.66 | |
| Feature ConcatEvaluation Protocol=Leave-One-Video-Out (LOVO)2026.04 | 71.83 | 35.35 | 66.66 | |
| Context-onlyEvaluation Protocol=Leave-One-Video-Out (LOVO)2026.04 | 60.87 | 48.8 | 71.17 | |
| PoE-Ctx+PoseEvaluation Protocol=Leave-One-Video-Out (LOVO)2026.04 | 60.55 | 43.56 | 67.29 | |
| Audio-onlyEvaluation Protocol=Leave-One-Video-Out (LOVO)2026.04 | 50.87 | 38.76 | 49.77 | |
| Pose-onlyEvaluation Protocol=Leave-One-Video-Out (LOVO)2026.04 | 39.36 | 36.25 | 36.23 |