mlsanityData to 14 Oct 2026

3D Indoor Instance Segmentation in an Open-World

arXiv 2309.14338 ↗First posted 25 Sep 2023cs.CVcited by 7

Not on the map. Why?

Abstract (arXiv)

Existing 3D instance segmentation methods typically assume that all semantic classes to be segmented would be available during training and only seen categories are segmented at inference. We argue that such a closed-world assumption is restrictive and explore for the first time 3D indoor instance segmentation in an open-world setting, where the model is allowed to distinguish a set of known classes as well as identify an unknown object as unknown and then later incrementally learning the semantic category of the unknown when the corresponding category labels are available. To this end, we introduce an open-world 3D indoor instance segmentation method, where an auto-labeling scheme is employed to produce pseudo-labels during training and induce separation to separate known and unknown category labels. We further improve the pseudo-labels quality at inference by adjusting the unknown class probability based on the objectness score distribution. We also introduce carefully curated open-world splits leveraging realistic scenarios based on inherent object distribution, region-based indoor scene exploration and randomness aspect of open-world classes. Extensive experiments reveal the efficacy of the proposed contributions leading to promising open-world 3D instance segmentation performance.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Builds on 38 of its 42 references that are in the map. First posted on arXiv 25 Sep 2023.

Builds on

38 of 42 references.

References · 38

  1. 2017Attention Is All You Needcited by 60,361
  2. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
  3. 2021Masked-attention Mask Transformer for Universal Image Segmentationcited by 2,351
  4. 2021Per-Pixel Classification is Not All You Need for Semantic Segmentationcited by 1,152
  5. 2019Large-Scale Long-Tailed Recognition in an Open Worldcited by 705
  6. 2020PointContrast: Unsupervised Pre-training for 3D Point Cloud Understandingcited by 504
  7. 2016Visual Relationship Detection with Language Priorscited by 334
  8. 2020PointGroup: Dual-Set Point Grouping for 3D Instance Segmentationcited by 321
  9. 2022Mask3D: Mask Transformer for 3D Semantic Instance Segmentationcited by 303
  10. 2021Towards Open World Object Detectioncited by 297
  11. 2022Language-Grounded Indoor 3D Semantic Segmentation in the Wildcited by 246
  12. 2014Towards Open World Recognitioncited by 226
26 more
  1. 20183D-SIS: 3D Semantic Instance Segmentation of RGB-D Scanscited by 217
  2. 2020Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contextscited by 210
  3. 2018Zero-Shot Object Detectioncited by 189
  4. 2020OccuSeg: Occupancy-aware 3D Instance Segmentationcited by 177
  5. 2017SGPN: Similarity Group Proposal Network for 3D Point Cloud Instance Segmentationcited by 169
  6. 2019Learning Object Bounding Boxes for 3D Instance Segmentation on Point Cloudscited by 156
  7. 2018GSPN: Generative Shape Proposal Network for 3D Instance Segmentation in Point Cloudcited by 146
  8. 20203D-MPA: Multi Proposal Aggregation for 3D Semantic Instance Segmentationcited by 141
  9. 2021Unidentified Video Objects: A Benchmark for Dense, Open-World Segmentationcited by 133
  10. 2021OW-DETR: Open-world Detection Transformercited by 129
  11. 2021Hierarchical Aggregation for 3D Instance Segmentationcited by 127
  12. 2022Superpoint Transformer for 3D Scene Instance Segmentationcited by 116
  13. 2017Dropout Sampling for Robust Object Detection in Open-Set Conditionscited by 112
  14. 2021Learning Open-World Object Proposals without Learning to Classifycited by 109
  15. 2021Instance Segmentation in 3D Scenes using Semantic Superpoint Tree Networkscited by 99
  16. 2022Semantic Abstraction: Open-World 3D Scene Understanding from 2D Vision-Language Modelscited by 95
  17. 2023CLIP-FO3D: Learning Free Open-world 3D Scene Representations from 2D Dense CLIPcited by 89
  18. 20193D Instance Segmentation via Multi-Task Metric Learningcited by 80
  19. 2020DyCo3D: Robust Instance Segmentation of 3D Point Clouds through Dynamic Convolutioncited by 75
  20. 2022PROB: Probabilistic Objectness for Open World Object Detectioncited by 57
  21. 2019Point Cloud Instance Segmentation using Probabilistic Embeddingscited by 52
  22. 2022Open-World Instance Segmentation: Exploiting Pseudo Ground Truth From Learned Pairwise Affinitycited by 41
  23. 2021Learning to Detect Every Thing in an Open Worldcited by 37
  24. 2020Learning Gaussian Instance Segmentation in Point Cloudscited by 33
  25. 2022Open-world Semantic Segmentation for LIDAR Point Cloudscited by 29
  26. 2023CAT: LoCalization and IdentificAtion Cascade Detection Transformer for Open-World Object Detectioncited by 28

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 2; 2025: 3; 2026: 2.

7 papers.

  1. 2024Open-YOLO 3D: Towards Fast and Accurate Open-Vocabulary 3D Instance Segmentationcited by 30
  2. 2024Shape2Scene: 3D Scene Representation Learning Through Pre-training on Shape Datacited by 8
  3. 2025Towards 3D Objectness Learning in an Open Worldcited by 2
  4. 2025CLIMB-3D: Continual Learning for Imbalanced 3D Instance Segmentationcited by 1
  5. 2025A Novel Decomposed Feature-Oriented Framework for Open-Set Semantic Segmentation on LiDAR Datacited by 1
  6. 2026SCOPE: Scene-Contextualized Incremental Few-Shot 3D Segmentationcited by 1
  7. 2026ClickSeg3D: Few-Click Interactive Segmentation via Semantic Embeddingscited by 0