mlsanityData to 14 Oct 2026

Feed-Forward SceneDINO for Unsupervised Semantic Scene Completion

arXiv 2507.06230 ↗First posted 8 Jul 2025cs.CVcited by 13

Not on the map. Why?

Abstract (arXiv)

Semantic scene completion (SSC) aims to infer both the 3D geometry and semantics of a scene from single images. In contrast to prior work on SSC that heavily relies on expensive ground-truth annotations, we approach SSC in an unsupervised setting. Our novel method, SceneDINO, adapts techniques from self-supervised representation learning and 2D unsupervised scene understanding to SSC. Our training exclusively utilizes multi-view consistency self-supervision without any form of semantic or geometric ground truth. Given a single input image, SceneDINO infers the 3D geometry and expressive 3D DINO features in a feed-forward manner. Through a novel 3D feature distillation approach, we obtain unsupervised 3D semantics. In both 3D and 2D unsupervised scene understanding, SceneDINO reaches state-of-the-art segmentation accuracy. Linear probing our 3D features matches the segmentation accuracy of a current supervised SSC approach. Additionally, we showcase the domain generalization and multi-view consistency of SceneDINO, taking the first steps towards a strong foundation for single image 3D scene understanding.

Key facts. Cited by 13 papers in the map (data through 14 Oct 2026). Builds on 101 of its 129 references that are in the map. First posted on arXiv 8 Jul 2025.

Builds on

101 of 129 references.

References · 101

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  3. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  4. 2023Segment Anythingcited by 8,234
  5. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  6. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,513
  7. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  8. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  9. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,750
  10. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  11. 2024SAM 2: Segment Anything in Images and Videoscited by 2,863
  12. 2020Unsupervised Learning of Visual Features by Contrasting Cluster Assignmentscited by 2,857
38 more
  1. 2020RAFT: Recurrent All-Pairs Field Transforms for Optical Flowcited by 2,415
  2. 2020Improved Baselines with Momentum Contrastive Learningcited by 2,287
  3. 2021Vision Transformers for Dense Predictioncited by 1,803
  4. 2021An Empirical Study of Training Self-Supervised Vision Transformerscited by 1,616
  5. 2020pixelNeRF: Neural Radiance Fields from One or Few Imagescited by 1,584
  6. 20163D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotationcited by 1,476
  7. 2016Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzlescited by 1,382
  8. 2018BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learningcited by 1,288
  9. 2018Learning deep representations by mutual information estimation and maximizationcited by 1,273
  10. 2015Unsupervised Visual Representation Learning by Context Predictioncited by 1,066
  11. 2018Deep Clustering for Unsupervised Learning of Visual Featurescited by 1,028
  12. 2020ORB-SLAM3: An Accurate Open-Source Library for Visual, Visual-Inertial and Multi-Map SLAMcited by 878
  13. 2016Unsupervised Monocular Depth Estimation with Left-Right Consistencycited by 820
  14. 2019Learning Representations by Maximizing Mutual Information Across Viewscited by 734
  15. 2019Data-Efficient Image Recognition with Contrastive Predictive Codingcited by 673
  16. 2020Prototypical Contrastive Learning of Unsupervised Representationscited by 640
  17. 2021Masked Feature Prediction for Self-Supervised Visual Pre-Trainingcited by 576
  18. 2021KITTI-360: A Novel Dataset and Benchmarks for Urban Scene Understanding in 2D and 3Dcited by 575
  19. 2022OpenScene: 3D Scene Understanding with Open Vocabulariescited by 559
  20. 2023LERF: Language Embedded Radiance Fieldscited by 543
  21. 2021Neural Fields in Visual Computing and Beyondcited by 540
  22. 2019Self-labelling via simultaneous clustering and representation learningcited by 528
  23. 2016Semantic Scene Completion from a Single Depth Imagecited by 472
  24. 2018Invariant Information Clustering for Unsupervised Image Classification and Segmentationcited by 455
  25. 2023Tri-Perspective View for Vision-Based 3D Semantic Occupancy Predictioncited by 409
  26. 2024Metric3Dv2: A Versatile Monocular Geometric Foundation Model for Zero-shot Metric Depth and Surface Normal Estimationcited by 395
  27. 2022Decomposing NeRF for Editing via Feature Field Distillationcited by 371
  28. 2021MonoScene: Monocular 3D Semantic Scene Completioncited by 363
  29. 2023VoxFormer: Sparse Voxel Transformer for Camera-based 3D Semantic Scene Completioncited by 313
  30. 2023OccFormer: Dual-path Transformer for Vision-based 3D Semantic Occupancy Predictioncited by 293
  31. 2023OpenMask3D: Open-Vocabulary 3D Instance Segmentationcited by 289
  32. 2022Unsupervised Semantic Segmentation by Distilling Feature Correspondencescited by 241
  33. 2020Sparse Single Sweep LiDAR Point Cloud Segmentation via Learning Contextual Shape Priors from Scene Completioncited by 223
  34. 2023EmerNeRF: Emergent Spatial-Temporal Scene Decomposition via Self-Supervisioncited by 217
  35. 2023Scene as Occupancycited by 209
  36. 2022Neural Feature Fusion Fields: 3D Distillation of Self-Supervised 2D Image Representationscited by 196
  37. 2020LMSCNet: Lightweight Multiscale 3D Semantic Completioncited by 188
  38. 2023Cut and Learn for Unsupervised Object Detection and Instance Segmentationcited by 187

50 most cited of 101 All 101 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 11.

13 papers.

Sort:

By area:

  1. 2026INSID3: Training-Free In-Context Segmentation with DINOv3cited by 12
  2. 2025OneOcc: Semantic Occupancy Prediction for Legged Robots with a Single Panoramic Cameracited by 8
  3. 2026OccAny: Generalized Unconstrained Urban 3D Occupancycited by 6
  4. 2025OccuFly: A 3D Vision Benchmark for Semantic Scene Completion from the Aerial Perspectivecited by 5
  5. 2026Control-DINO: Feature Space Conditioning for Controllable Image-to-Video Diffusioncited by 1
  6. 2026Scene-Centric Unsupervised Video Panoptic Segmentationcited by 1
  7. 2026FLM-Occ: Feed-forward Likelihood Maximization for Efficient Indoor Occupancy Predictioncited by 1
  8. 2026Rebenchmarking Unsupervised Monocular 3D Occupancy Predictioncited by 0
  9. 2026The Neural Compass: Probabilistic Relative Feature Fields for Robotic Searchcited by 0
  10. 2026ReconSplat: Generalizable 3D Scene Reconstruction Beyond Observed Viewscited by 0
  11. 2026TriO: Tri-Modal Unsupervised Occupancy World Model for Anything Perceptioncited by 0
  12. 2026CrowdOcc: Monocular Semantic Scene Completion for Quadruped Robots in Crowded Indoor Environmentscited by 0
1 more
  1. 2026GNFormer: Gaussian Neural Field Transformer for Efficient Camera-based Occupancy Predictioncited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026SceneBench: A Hierarchical Benchmark for Vision-Language Understanding of 3D Scenes
  2. 2024A Unified Framework for 3D Scene Understanding
  3. 2023CLIP-FO3D: Learning Free Open-world 3D Scene Representations from 2D Dense CLIP
  4. 2023SCPNet: Semantic Scene Completion on Point Cloud
  5. 2026SceneScaffold: Active Scene-State Construction for Unified 3D Scene Understanding
  6. 2026Beyond Isolated Objects: Relationship-aware Open Vocabulary Scene Understanding via 3D Scene Graph Analysis
  7. 2025Global-Aware Monocular Semantic Scene Completion with State Space Models
  8. 2023Bridging the Domain Gap: Self-Supervised 3D Scene Understanding with Foundation Models