3D Object Tracking on V2X-Seq
44.4AMOTA (0-150m Overall)Long-SCOPE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Long-SCOPEBackbone=ResNet50, Number of Queries=900 + 40 (max), Iterative Refinement Layers=42026.04 | 44.4 | 57.6 | 36.1 | 5.9 | |
| Long-SCOPE w/o GQGBackbone=ResNet50, Number of Queries=900, Iterative Refinement Layers=42026.04 | 41.6 | 55.9 | 34.1 | 4.6 | |
| Long-SCOPE w/o CAABackbone=ResNet50, Number of Queries=900 + 40 (max)2026.04 | 38.2 | 58.8 | 28 | 0.5 | |
| V2X-ViTBackbone=BEVFormer, Grid Resolution=1m x 1m2026.04 | 34.5 | 50.9 | 25.1 | 2.7 | |
| SparseCoopBackbone=ResNet50, Number of Queries=9002026.04 | 32.8 | 53.4 | 23.2 | 0 | |
| Late Fusion BaselineBackbone=ResNet502026.04 | 27.1 | 37.5 | 18.5 | 3.2 | |
| CoopTrackBackbone=ResNet50, Number of Queries=9002026.04 | 17.1 | 25.5 | 8.8 | 0.6 | |
| UniV2XBackbone=ResNet50, Number of Queries=9002026.04 | 1.9 | 10.1 | 0 | 0 | |
| Where2CommBackbone=BEVFormer, Grid Resolution=1m x 1m2026.04 | 0.4 | 8 | 0 | 0 |