mlsanityData to 14 Oct 2026

Once Detected, Never Lost: Surpassing Human Performance in Offline LiDAR based 3D Object Detection

arXiv 2304.12315 ↗First posted 24 Apr 2023cs.CV cs.ROcited by 28

Not on the map. Why?

Abstract (arXiv)

This paper aims for high-performance offline LiDAR-based 3D object detection. We first observe that experienced human annotators annotate objects from a track-centric perspective. They first label the objects with clear shapes in a track, and then leverage the temporal coherence to infer the annotations of obscure objects. Drawing inspiration from this, we propose a high-performance offline detector in a track-centric perspective instead of the conventional object-centric perspective. Our method features a bidirectional tracking module and a track-centric learning module. Such a design allows our detector to infer and refine a complete track once the object is detected at a certain moment. We refer to this characteristic as "onCe detecTed, neveR Lost" and name the proposed system CTRL. Extensive experiments demonstrate the remarkable performance of our method, surpassing the human-level annotating accuracy and the previous state-of-the-art methods in the highly competitive Waymo Open Dataset without model ensemble. The code will be made publicly available at https://github.com/tusen-ai/SST.

Key facts. Cited by 28 papers in the map (data through 14 Oct 2026). Builds on 53 of its 61 references that are in the map. First posted on arXiv 24 Apr 2023.

Builds on

53 of 61 references.

References · 53

  1. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  2. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,401
  3. 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,451
  4. 2018PointPillars: Fast Encoders for Object Detection from Point Cloudscited by 1,715
  5. 2022BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformerscited by 1,218
  6. 2020Center-based 3D Object Detection and Trackingcited by 1,144
  7. 2018PointRCNN: 3D Object Proposal Generation and Detection from Point Cloudcited by 941
  8. 2019PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detectioncited by 891
  9. 2022BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representationcited by 850
  10. 2021BEVDet: High-performance Multi-camera 3D Object Detection in Bird-Eye-Viewcited by 713
  11. 2021DETR3D: 3D Object Detection from Multi-view Images via 3D-to-2D Queriescited by 658
  12. 2019Deep Hough Voting for 3D Object Detection in Point Cloudscited by 648
38 more
  1. 2022PETR: Position Embedding Transformation for Multi-View 3D Object Detectioncited by 552
  2. 20203DSSD: Point-based 3D Single Stage Object Detectorcited by 440
  3. 2020Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detectioncited by 409
  4. 2019From Points to Parts: 3D Object Detection from Point Cloud with Part-aware and Part-aggregation Networkcited by 368
  5. 2021PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detectioncited by 232
  6. 2023VoxelNeXt: Fully Sparse VoxelNet for 3D Object Detection and Trackingcited by 222
  7. 2021Embracing Single Stride 3D Object Detector with Sparse Transformercited by 193
  8. 2021Offboard 3D Object Detection from Point Cloud Sequencescited by 145
  9. 2021RangeDet:In Defense of Range View for LiDAR-based 3D Object Detectioncited by 140
  10. 2019LaserNet: An Efficient Probabilistic 3D Object Detector for Autonomous Drivingcited by 133
  11. 2023DSVT: Dynamic Sparse Voxel Transformer with Rotated Setscited by 122
  12. 2021RSN: Range Sparse Net for Efficient, Accurate LiDAR 3D Object Detectioncited by 115
  13. 2020GNN3DMOT: Graph Neural Network for 3D Multi-Object Tracking with Multi-Feature Learningcited by 113
  14. 2021SimpleTrack: Understanding and Rethinking 3D Multi-object Trackingcited by 103
  15. 2022CenterFormer: Center-based Transformer for 3D Object Detectioncited by 99
  16. 2017Annotating Object Instances with a Polygon-RNNcited by 88
  17. 2021AFDetV2: Rethinking the Necessity of the Second Stage for Object Detection from Point Cloudscited by 88
  18. 2022Fully Sparse 3D Object Detectioncited by 84
  19. 2023LoGoNet: Towards Accurate 3D Object Detection with Local-to-Global Cross-Modal Fusioncited by 80
  20. 2020Range Conditioned Dilated Convolutions for Scale Invariant 3D Object Detectioncited by 78
  21. 2019Autolabeling 3D Objects with Differentiable Rendering of SDF Shape Priorscited by 74
  22. 2020Probabilistic 3D Multi-Modal, Multi-Object Tracking for Autonomous Drivingcited by 73
  23. 20213D-MAN: 3D Multi-frame Attention Network for Object Detectioncited by 73
  24. 2023FlatFormer: Flattened Window Attention for Efficient Point Cloud Transformercited by 73
  25. 2020Weakly Supervised 3D Object Detection from Lidar Point Cloudcited by 67
  26. 2022MPPNet: Multi-Frame Feature Intertwining with Proxy Points for 3D Temporal Object Detectioncited by 65
  27. 2022GD-MAE: Generative Decoder for MAE Pre-training on LiDAR Point Cloudscited by 57
  28. 2021Exploring Simple 3D Multi-Object Tracking for Autonomous Drivingcited by 56
  29. 20193D Multi-Object Tracking: A Baseline and New Evaluation Metricscited by 47
  30. 2019Deep Active Learning for Efficient Training of a LiDAR 3D Object Detectorcited by 45
  31. 2021Auto4D: Learning to Label 4D Objects from Sequential Point Cloudscited by 42
  32. 2021Immortal Tracker: Tracklet Never Diescited by 42
  33. 2022Graph R-CNN: Towards Accurate 3D Object Detection with Semantic-Decorated Local Graphcited by 42
  34. 2023Super Sparse 3D Object Detectioncited by 32
  35. 2021Model-free Vehicle Tracking and State Estimation in Point Cloud Sequencescited by 25
  36. 2022ConQueR: Query Contrast Voxel-DETR for 3D Object Detectioncited by 21
  37. 2022INT: Towards Infinite-frames 3D Detection with An Efficient Frameworkcited by 14
  38. 2018Leveraging Pre-Trained 3D Object Detection Models For Fast Ground Truth Generationcited by 12

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 5; 2024: 10; 2025: 8; 2026: 5.

28 papers.

Sort:

By area:

  1. 2024Voxel Mamba: Group-Free State Space Models for Point Cloud based 3D Object Detectioncited by 66
  2. 2023FSD V2: Improving Fully Sparse 3D Object Detection with Virtual Voxelscited by 34
  3. 2025Perspective-Invariant 3D Object Detectioncited by 18
  4. 2024MCTrack: A Unified 3D Multi-Object Tracking Framework for Autonomous Drivingcited by 17
  5. 2023MS3D++: Ensemble of Experts for Multi-Source Unsupervised Domain Adaptation in 3D Object Detectioncited by 14
  6. 2024Offboard Occupancy Refinement with Hybrid Propagation for Autonomous Drivingcited by 12
  7. 20253EED: Ground Everything Everywhere in 3Dcited by 12
  8. 2024MixSup: Mixed-grained Supervision for Label-efficient LiDAR-based 3D Object Detectioncited by 9
  9. 2024RobMOT: Robust 3D Multi-Object Tracking by Observational Noise and State Estimation Drift Mitigation on LiDAR PointCloudcited by 9
  10. 2024SOAP: Cross-sensor Domain Adaptation for 3D Object Detection Using Stationary Object Aggregation Pseudo-labellingcited by 8
  11. 2023Offline Tracking with Object Permanencecited by 7
  12. 2023LabelFormer: Object Trajectory Refinement for Offboard Perception from LiDAR Point Cloudscited by 7
16 more
  1. 2024Towards Stable 3D Object Detectioncited by 7
  2. 2025Learning to Detect Objects from Multi-Agent LiDAR Scans without Manual Labelscited by 7
  3. 2024ZOPP: A Framework of Zero-shot Offboard Panoptic Perception for Autonomous Drivingcited by 6
  4. 2024Towards Flexible 3D Perception: Object-Centric Occupancy Completion Augments 3D Object Detectioncited by 3
  5. 2025Learning better representations for crowded pedestrians in offboard LiDAR-camera 3D tracking-by-detectioncited by 2
  6. 2026Offline-Poly: A Polyhedral Framework For Offline 3D Multi-Object Trackingcited by 2
  7. 2025Towards Accurate State Estimation: Motion Dynamics Kalman Filter for 3D Multi-Object Trackingcited by 1
  8. 2025M^3Detection: Multi-Frame Multi-Level Feature Fusion for Multi-Modal 3D Object Detection with Camera and 4D Imaging Radarcited by 1
  9. 2023RBF Weighted Hyper-Involution for RGB-D Object Detectioncited by 0
  10. 2024RegTrack: Simplicity Beneath Complexity in Robust Multi-Modal 3D Multi-Object Trackingcited by 0
  11. 2025From Binary to Semantic: Utilizing Large-Scale Binary Occupancy Data for 3D Semantic Occupancy Predictioncited by 0
  12. 2025MEGA-PCC: A Mamba-based Efficient Approach for Joint Geometry and Attribute Point Cloud Compressioncited by 0
  13. 2026UniDA3D: A Unified Domain-Adaptive Framework for Multi-View 3D Object Detectioncited by 0
  14. 2026MotionSync: Non-Causal Refinement of Causal Tracker for Label-Efficient 3D Perceptioncited by 0
  15. 2026Do Spinning Radar Doppler Velocity Measurements Improve Vehicle Detection and Tracking?cited by 0
  16. 2026A Differentiable Optimization Framework for Registering Sequential Bounding Boxes with Point Cloud Streamcited by 0