MVTrackTrans uses view-ground interactions in a Transformer to improve multi-view crowd tracking on large scenes and outperforms prior methods on two newly collected large datasets.
Deformable detr: Deformable transformers for end-to-end object detection
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
OptiMVMap selects optimal helper vehicles and applies cross-vehicle attention with noise filtering to fuse multi-vehicle views, improving vectorized map accuracy by over 10 mAP on nuScenes compared to MapTRv2.
citing papers explorer
-
Multi-view Crowd Tracking Transformer with View-Ground Interactions Under Large Real-World Scenes
MVTrackTrans uses view-ground interactions in a Transformer to improve multi-view crowd tracking on large scenes and outperforms prior methods on two newly collected large datasets.
-
OptiMVMap: Offline Vectorized Map Construction via Optimal Multi-vehicle Perspectives
OptiMVMap selects optimal helper vehicles and applies cross-vehicle attention with noise filtering to fuse multi-vehicle views, improving vectorized map accuracy by over 10 mAP on nuScenes compared to MapTRv2.