mlsanityData to 14 Oct 2026

LEGO: Learning and Graph-Optimized Modular Tracker for Online Multi-Object Tracking with Point Clouds

arXiv 2308.09908 ↗First posted 19 Aug 2023cs.CV cs.AIcited by 9

Not on the map. Why?

Abstract (arXiv)

Online multi-object tracking (MOT) plays a pivotal role in autonomous systems. The state-of-the-art approaches usually employ a tracking-by-detection method, and data association plays a critical role. This paper proposes a learning and graph-optimized (LEGO) modular tracker to improve data association performance in the existing literature. The proposed LEGO tracker integrates graph optimization and self-attention mechanisms, which efficiently formulate the association score map, facilitating the accurate and efficient matching of objects across time frames. To further enhance the state update process, the Kalman filter is added to ensure consistent tracking by incorporating temporal coherence in the object states. Our proposed method utilizing LiDAR alone has shown exceptional performance compared to other online tracking approaches, including LiDAR-based and LiDAR-camera fusion-based methods. LEGO ranked 1st at the time of submitting results to KITTI object tracking evaluation ranking board and remains 2nd at the time of submitting this paper, among all online trackers in the KITTI MOT benchmark for cars1

Key facts. Cited by 9 papers in the map (data through 14 Oct 2026). Builds on 47 of its 104 references that are in the map. First posted on arXiv 19 Aug 2023.

Builds on

47 of 104 references.

References · 47

  1. 2017Graph Attention Networkscited by 8,923
  2. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  3. 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,451
  4. 2020Center-based 3D Object Detection and Trackingcited by 1,144
  5. 2018PointRCNN: 3D Object Proposal Generation and Detection from Point Cloudcited by 941
  6. 2021ByteTrack: Multi-Object Tracking by Associating Every Detection Boxcited by 823
  7. 2020HOTA: A Higher Order Metric for Evaluating Multi-Object Trackingcited by 552
  8. 2020FairMOT: On the Fairness of Detection and Re-Identification in Multiple Object Trackingcited by 490
  9. 2020Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloudcited by 335
  10. 2022StrongSORT: Make DeepSORT Great Againcited by 213
  11. 2020GNN3DMOT: Graph Neural Network for 3D Multi-Object Tracking with Multi-Feature Learningcited by 113
  12. 2021SimpleTrack: Understanding and Rethinking 3D Multi-object Trackingcited by 103
35 more
  1. 2021EagerMOT: 3D Multi-Object Tracking via Sensor Fusioncited by 83
  2. 2018Joint Monocular 3D Vehicle Detection and Trackingcited by 77
  3. 2020Probabilistic 3D Multi-Modal, Multi-Object Tracking for Autonomous Drivingcited by 73
  4. 2020Probabilistic 3D Multi-Object Tracking for Autonomous Drivingcited by 64
  5. 2023LXL: LiDAR Excluded Lean 3D Object Detection with 4D Imaging Radar and Camera Fusioncited by 58
  6. 2019Robust Multi-Modality Multi-Object Trackingcited by 56
  7. 2021Exploring Simple 3D Multi-Object Tracking for Autonomous Drivingcited by 56
  8. 2022CAMO-MOT: Combined Appearance-Motion Optimization for 3D Multi-Object Tracking with Camera-LiDAR Fusioncited by 50
  9. 2023SMURF: Spatial Multi-Representation Fusion for 3D Object Detection with 4D Imaging Radarcited by 49
  10. 2021Learnable Graph Matching: Incorporating Graph Partitioning with Deep Feature Learning for Multiple Object Trackingcited by 47
  11. 2021Learning a Proposal Classifier for Multiple Object Trackingcited by 42
  12. 2021Immortal Tracker: Tracklet Never Diescited by 42
  13. 2021Learnable Online Graph Representations for 3D Multi-Object Trackingcited by 38
  14. 2022DeepFusionMOT: A 3D Multi-Object Tracking Framework Based on Camera-LiDAR Fusion with Deep Associationcited by 33
  15. 2021Score refinement for confidence-based 3D multi-object trackingcited by 32
  16. 2022PolarMOT: How Far Can Geometric Relations Take Us in 3D Multi-Object Tracking?cited by 29
  17. 2019FANTrack: 3D Multi-Object Tracking with Feature Association Networkcited by 28
  18. 20223D Multi-Object Tracking Using Graph Neural Networks with Cross-Edge Modality Attentioncited by 27
  19. 2021Joint Multi-Object Detection and Tracking with Camera-LiDAR Fusion for Autonomous Drivingcited by 26
  20. 2022RaLiBEV: Radar and LiDAR BEV Fusion Learning for Anchor Box Free Object Detection Systemscited by 23
  21. 2021Deep Instance Segmentation with Automotive Radar Detection Pointscited by 22
  22. 2022ShaSTA: Modeling Shape and Spatio-Temporal Affinities for 3D Multi-Object Trackingcited by 21
  23. 20233DMOTFormer: Graph Transformer for Online 3D Multi-Object Trackingcited by 21
  24. 2023ByteTrackV2: 2D and 3D Multi-Object Tracking by Associating Every Detection Boxcited by 20
  25. 2019Graph Neural Based End-to-end Data Association Framework for Online Multiple-Object Trackingcited by 15
  26. 2022GNN-PMB: A Simple but Effective Online 3D Multi-Object Tracker without Bells and Whistlescited by 15
  27. 2022Transformers for Multi-Object Tracking on Point Cloudscited by 14
  28. 2022Contrastive Learning for Automotive mmWave Radar Detection Points Based Instance Segmentationcited by 12
  29. 2022Neural Enhanced Belief Propagation for Multiobject Trackingcited by 11
  30. 2022Minkowski Tracker: A Sparse Spatio-Temporal R-CNN for Joint Object Detection and Trackingcited by 9
  31. 2022InterTrack: Interaction Transformer for 3D Multi-Object Trackingcited by 6
  32. 20223DMODT: Attention-Guided Affinities for Joint Detection & Tracking in 3D Point Cloudscited by 6
  33. 2020A CRF-based Framework for Tracklet Inactivation in Online Multi-Object Trackingcited by 4
  34. 20233D Multi-Object Tracking Based on Uncertainty-Guided Data Associationcited by 3
  35. 2023Trajectory Poisson multi-Bernoulli mixture filter for traffic monitoring using a dronecited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 3; 2025: 5; 2026: 1.

9 papers.

  1. 2025Doracamom: Joint 3D Detection and Occupancy Prediction with Multi-view 4D Radars and Cameras for Omnidirectional Perceptioncited by 22
  2. 2024MCTrack: A Unified 3D Multi-Object Tracking Framework for Autonomous Drivingcited by 17
  3. 2024On the Federated Learning Framework for Cooperative Perceptioncited by 2
  4. 2025LiDAR MOT-DETR: A LiDAR-based Two-Stage Transformer for 3D Multiple Object Trackingcited by 2
  5. 2024One Homography is All You Need: IMM-based Joint Homography and Multiple Object State Estimationcited by 1
  6. 2025OptiPMB: Enhancing 3D Multi-Object Tracking with Optimized Poisson Multi-Bernoulli Filteringcited by 1
  7. 2025Stable at Any Speed: Speed-Driven Multi-Object Tracking with Learnable Kalman Filteringcited by 1
  8. 2025FutrTrack: A Camera-LiDAR Fusion Transformer for 3D Multiple Object Trackingcited by 0
  9. 2026MotionSync: Non-Causal Refinement of Causal Tracker for Label-Efficient 3D Perceptioncited by 0