mlsanityData to 7 Oct 2026

MASAR: Motion-Appearance Synergy Refinement for Joint Detection and Trajectory Forecasting

arXiv 2602.13003 ↗First posted 13 Feb 2026cs.CV cs.LGcited by 1

Not on the map. Why?

Abstract (arXiv)

Classical autonomous driving systems connect perception and prediction modules via hand-crafted bounding-box interfaces, limiting information flow and propagating errors to downstream tasks. Recent research aims to develop end-to-end models that jointly address perception and prediction; however, they often fail to fully exploit the synergy between appearance and motion cues, relying mainly on short-term visual features. We follow the idea of "looking backward to look forward", and propose MASAR, a novel fully differentiable framework for joint 3D detection and trajectory forecasting compatible with any transformer-based 3D detector. MASAR employs an object-centric spatio-temporal mechanism that jointly encodes appearance and motion features. By predicting past trajectories and refining them using guidance from appearance cues, MASAR captures long-term temporal dependencies that enhance future trajectory forecasting. Experiments conducted on the nuScenes dataset demonstrate MASAR's effectiveness, showing improvements of over 20% in minADE and minFDE while maintaining robust detection performance. Code and models are available at https://github.com/aminmed/MASAR.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 40 of its 44 references that are in the map. First posted on arXiv 13 Feb 2026.

Builds on

40 of 44 references.

References · 40

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2016Feature Pyramid Networks for Object Detectioncited by 5,021
  4. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,485
  5. 2020Deformable DETR: Deformable Transformers for End-to-End Object Detectioncited by 3,410
  6. 2022BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformerscited by 1,213
  7. 2022Planning-oriented Autonomous Drivingcited by 1,068
  8. 2020Lift, Splat, Shoot: Encoding Images From Arbitrary Camera Rigs by Implicitly Unprojecting to 3Dcited by 1,040
  9. 2021DETR3D: 3D Object Detection from Multi-view Images via 3D-to-2D Queriescited by 655
  10. 2022PETR: Position Embedding Transformation for Multi-View 3D Object Detectioncited by 550
  11. 2020VectorNet: Encoding HD Maps and Agent Dynamics from Vectorized Representationcited by 528
  12. 2019MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Predictioncited by 417
28 more
  1. 2018ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worstcited by 391
  2. 2022BEVDet4D: Exploit Temporal Cues in Multi-camera 3D Object Detectioncited by 340
  3. 2021Categorical Depth Distribution Network for Monocular 3D Object Detectioncited by 326
  4. 2023Exploring Object-Centric Temporal Modeling for Efficient Multi-View 3D Object Detectioncited by 313
  5. 2022Motion Transformer with Global Intention Localization and Local Movement Refinementcited by 267
  6. 2021FIERY: Future Instance Prediction in Bird's-Eye View from Surround Monocular Camerascited by 239
  7. 2024SparseDrive: End-to-End Autonomous Driving via Sparse Scene Representationcited by 239
  8. 2020Fast and Furious: Real Time End-to-End 3D Detection, Tracking and Motion Forecasting with a Single Convolutional Netcited by 228
  9. 2022BEVerse: Unified Perception and Prediction in Birds-Eye-View for Vision-Centric Autonomous Drivingcited by 191
  10. 2021Scene Transformer: A unified architecture for predicting multiple agent trajectoriescited by 165
  11. 2022Sparse4D: Multi-view 3D Object Detection with Sparse Spatial-Temporal Fusioncited by 161
  12. 2023SparseBEV: High-Performance Sparse 3D Object Detection from Multi-Camera Videoscited by 149
  13. 2020PnPNet: End-to-End Perception and Prediction with Tracking in the Loopcited by 143
  14. 2018ROI-10D: Monocular Lifting of 2D Detection to 6D Pose and Metric Shapecited by 123
  15. 2022ViP3D: End-to-end Visual Trajectory Prediction via 3D Agent Queriescited by 116
  16. 2022MUTR3D: A Multi-camera Tracking Framework via 3D-to-2D Queriescited by 90
  17. 2022AdaMixer: A Fast-Converging Query-Based Object Detectorcited by 78
  18. 2023Sparse4D v3: Advancing End-to-End 3D Detection and Trackingcited by 69
  19. 2023Exploring Recurrent Long-term Temporal Fusion for Multi-view 3D Perceptioncited by 65
  20. 2023BEVNeXt: Reviving Dense BEV Frameworks for 3D Object Detectioncited by 35
  21. 2023TBP-Former: Learning Temporal Bird's-Eye-View Pyramid for Joint Perception and Prediction in Vision-Centric Autonomous Drivingcited by 30
  22. 2022Forecasting from LiDAR via Future Object Detectioncited by 28
  23. 2022Perceive, Interact, Predict: Learning Dynamic and Static Clues for End-to-End Motion Predictioncited by 28
  24. 2023Towards Motion Forecasting with Real-World Perception Inputs: Are End-to-End Approaches Competitive?cited by 22
  25. 2024DualAD: Disentangling the Dynamic and Static World for End-to-End Drivingcited by 17
  26. 2024DeTra: A Unified Model for Object Detection and Trajectory Forecastingcited by 15
  27. 2023Delving into Motion-Aware Matching for Monocular 3D Object Trackingcited by 11
  28. 2025ForeSight: Multi-View Streaming Joint Object Detection and Trajectory Forecastingcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026MC-DeTra: Motion-Consistent Joint Object Detection and Socially-Aware Trajectory Forecasting in Bird's-Eye-View Imagescited by 0