mlsanityData to 7 Oct 2026

Towards Flexible 3D Perception: Object-Centric Occupancy Completion Augments 3D Object Detection

arXiv 2412.05154 ↗First posted 6 Dec 2024cs.CV cs.AIcited by 3

Not on the map. Why?

Abstract (arXiv)

While 3D object bounding box (bbox) representation has been widely used in autonomous driving perception, it lacks the ability to capture the precise details of an object's intrinsic geometry. Recently, occupancy has emerged as a promising alternative for 3D scene perception. However, constructing a high-resolution occupancy map remains infeasible for large scenes due to computational constraints. Recognizing that foreground objects only occupy a small portion of the scene, we introduce object-centric occupancy as a supplement to object bboxes. This representation not only provides intricate details for detected objects but also enables higher voxel resolution in practical applications. We advance the development of object-centric occupancy perception from both data and algorithm perspectives. On the data side, we construct the first object-centric occupancy dataset from scratch using an automated pipeline. From the algorithmic standpoint, we introduce a novel object-centric occupancy completion network equipped with an implicit shape decoder that manages dynamic-size occupancy generation. This network accurately predicts the complete object-centric occupancy volume for inaccurate object proposals by leveraging temporal information from long sequences. Our method demonstrates robust performance in completing object shapes under noisy detection and tracking conditions. Additionally, we show that our occupancy features significantly enhance the detection results of state-of-the-art 3D object detectors, especially for incomplete or distant objects in the Waymo Open Dataset.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 37 of its 43 references that are in the map. First posted on arXiv 6 Dec 2024.

Builds on

37 of 43 references.

References · 37

  1. 2017Attention Is All You Needcited by 59,932
  2. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  3. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,485
  4. 2019DeepSDF: Learning Continuous Signed Distance Functions for Shape Representationcited by 2,682
  5. 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,436
  6. 2022BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformerscited by 1,213
  7. 2020Center-based 3D Object Detection and Trackingcited by 1,143
  8. 2018PointRCNN: 3D Object Proposal Generation and Detection from Point Cloudcited by 939
  9. 20173D Semantic Segmentation with Submanifold Sparse Convolutional Networkscited by 746
  10. 2016Semantic Scene Completion from a Single Depth Imagecited by 468
  11. 2022MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstructioncited by 467
  12. 2023Tri-Perspective View for Vision-Based 3D Semantic Occupancy Predictioncited by 404
25 more
  1. 2023Occ3D: A Large-Scale 3D Occupancy Prediction Benchmark for Autonomous Drivingcited by 325
  2. 2023SurroundOcc: Multi-Camera 3D Occupancy Prediction for Autonomous Drivingcited by 310
  3. 2021PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detectioncited by 232
  4. 2020Sparse Single Sweep LiDAR Point Cloud Segmentation via Learning Contextual Shape Priors from Scene Completioncited by 222
  5. 2023VoxelNeXt: Fully Sparse VoxelNet for 3D Object Detection and Trackingcited by 221
  6. 2023OpenOccupancy: A Large Scale Benchmark for Surrounding Semantic Occupancy Perceptioncited by 218
  7. 2023Scene as Occupancycited by 208
  8. 2021Offboard 3D Object Detection from Point Cloud Sequencescited by 144
  9. 2018Scan2CAD: Learning CAD Model Alignment in RGB-D Scanscited by 142
  10. 2023SelfOcc: Self-Supervised Vision-Based 3D Occupancy Predictioncited by 133
  11. 2021LiDAR R-CNN: An Efficient and Universal 3D Object Detectorcited by 129
  12. 2022CenterFormer: Center-based Transformer for 3D Object Detectioncited by 99
  13. 2022SWFormer: Sparse Window Transformer for 3D Object Detection in Point Cloudscited by 92
  14. 2022Fully Sparse 3D Object Detectioncited by 84
  15. 2023SCPNet: Semantic Scene Completion on Point Cloudcited by 80
  16. 20213D-MAN: 3D Multi-frame Attention Network for Object Detectioncited by 73
  17. 2022MPPNet: Multi-Frame Feature Intertwining with Proxy Points for 3D Temporal Object Detectioncited by 65
  18. 2020FroDO: From Detections to 3D Objectscited by 51
  19. 2023HEDNet: A Hierarchical Encoder-Decoder Network for 3D Object Detection in Point Cloudscited by 45
  20. 2021Immortal Tracker: Tracklet Never Diescited by 42
  21. 2023FSD V2: Improving Fully Sparse 3D Object Detection with Virtual Voxelscited by 34
  22. 2023Super Sparse 3D Object Detectioncited by 32
  23. 2023Once Detected, Never Lost: Surpassing Human Performance in Offline LiDAR based 3D Object Detectioncited by 28
  24. 2023MoDAR: Using Motion Forecasting for 3D Object Detection in Point Cloud Sequencescited by 13
  25. 2023Learning a Room with the Occ-SDF Hybrid: Signed Distance Function Mingled with Occupancy Aids Scene Representationcited by 10

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 2; 2026: 1.

3 papers.

  1. 2025LEGO-Motion: Learning-Enhanced Grids with Occupancy Instance Modeling for Class-Agnostic Motion Predictioncited by 6
  2. 2025Learning Generalizable Shape Completion with SIM(3) Equivariancecited by 2
  3. 2026Unified Primitive Proxies for Structured Shape Completioncited by 0