Feature Identification on FeatureID-3DS
31.4Acc@5Poly+ZO+MI + GO-LED-OL_ar
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Poly+ZO+MI + GO-LED-OL_arVLM Backbone=Video-LLaMA-13B2025.12 | 31.4 | 14.3 | 44.5 | 19.7 | |
| Poly+ZO+MI + GH-LED_arVLM Backbone=Video-LLaMA-13B2025.12 | 30.9 | 10.3 | 53.3 | 27.2 | |
| Poly+ZO+MI + GO-LED-OL_arVLM Backbone=Chat-UniVi-13B2025.12 | 27.4 | 11.4 | 40.3 | 15.3 | |
| Poly+ZO+MI + GH-LED_arVLM Backbone=Chat-UniVi-13B2025.12 | 26.8 | 7.8 | 44.4 | 21.5 | |
| Poly+ZO+MI + GH-LEDVLM Backbone=Video-LLaMA-13B2025.12 | 26.2 | 4.9 | 39.6 | 13.9 | |
| Poly+ZO+MI + GO-LED-OLVLM Backbone=Video-LLaMA-13B2025.12 | 25.9 | 4.2 | 37.1 | 12.4 | |
| Extended KalmanVLM Backbone=Video-LLaMA-13B2025.12 | 23.6 | 3.6 | 34.2 | 8.5 | |
| Poly+ZO+MIVLM Backbone=Video-LLaMA-13B2025.12 | 23.6 | 3.1 | 35.8 | 9.2 | |
| Poly+ZO+MI + GH-LEDVLM Backbone=Chat-UniVi-13B2025.12 | 22.9 | 4.1 | 33.3 | 10.8 | |
| Linear + SGDVLM Backbone=Video-LLaMA-13B2025.12 | 22.8 | 2.9 | 30.1 | 6.5 | |
| RBF NetworkVLM Backbone=Video-LLaMA-13B2025.12 | 22.2 | 2.6 | 33 | 7.3 | |
| Extended KalmanVLM Backbone=Chat-UniVi-13B2025.12 | 21.4 | 2.9 | 27.1 | 7.2 | |
| PIDVLM Backbone=Video-LLaMA-13B2025.12 | 21.1 | 2.1 | 32.8 | 7.3 | |
| Poly+ZO+MIVLM Backbone=Chat-UniVi-13B2025.12 | 20.8 | 2.6 | 27.7 | 7.8 | |
| VLM (no control)VLM Backbone=Video-LLaMA-13B2025.12 | 19.8 | 0.3 | 25.1 | 0.5 | |
| RBF NetworkVLM Backbone=Chat-UniVi-13B2025.12 | 18.9 | 0.6 | 28.5 | 5.5 | |
| Poly+ZO+MI + GO-LED-OLVLM Backbone=Chat-UniVi-13B2025.12 | 18.6 | 3.4 | 29.1 | 9.3 | |
| PIDVLM Backbone=Chat-UniVi-13B2025.12 | 18.1 | 1.6 | 28.1 | 6.5 | |
| Linear + SGDVLM Backbone=Chat-UniVi-13B2025.12 | 17.5 | 2.4 | 26 | 5.9 | |
| VLM (no control)VLM Backbone=Chat-UniVi-13B2025.12 | 14.8 | 0.3 | 19.6 | 0.3 |