mlsanityData to 7 Oct 2026

GeoRefine: Self-Supervised Online Depth Refinement for Accurate Dense Mapping

arXiv 2205.01656 ↗First posted 3 May 2022cs.CVcited by 8

Not on the map. Why?

Abstract (arXiv)

We present a robust and accurate depth refinement system, named GeoRefine, for geometrically-consistent dense mapping from monocular sequences. GeoRefine consists of three modules: a hybrid SLAM module using learning-based priors, an online depth refinement module leveraging self-supervision, and a global mapping module via TSDF fusion. The proposed system is online by design and achieves great robustness and accuracy via: (i) a robustified hybrid SLAM that incorporates learning-based optical flow and/or depth; (ii) self-supervised losses that leverage SLAM outputs and enforce long-term geometric consistency; (iii) careful system design that avoids degenerate cases in online depth refinement. We extensively evaluate GeoRefine on multiple public datasets and reach as low as 5% absolute relative depth errors.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 61 of its 82 references that are in the map. First posted on arXiv 3 May 2022.

Builds on

61 of 82 references.

References · 61

  1. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,088
  2. 2020RAFT: Recurrent All-Pairs Field Transforms for Optical Flowcited by 2,385
  3. 2021Vision Transformers for Dense Predictioncited by 1,789
  4. 2019Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfercited by 1,484
  5. 2014Depth Map Prediction from a Single Image using a Multi-Scale Deep Networkcited by 1,264
  6. 2015ORB-SLAM: a Versatile and Accurate Monocular SLAM Systemcited by 983
  7. 2020ORB-SLAM3: An Accurate Open-Source Library for Visual, Visual-Inertial and Multi-Map SLAMcited by 862
  8. 2016Unsupervised Monocular Depth Estimation with Left-Right Consistencycited by 819
  9. 2016ORB-SLAM2: an Open-Source SLAM System for Monocular, Stereo and RGB-D Camerascited by 801
  10. 2017Unsupervised Learning of Depth and Ego-Motion from Videocited by 778
  11. 2018MegaDepth: Learning Single-View Depth Prediction from Internet Photoscited by 731
  12. 2018Deep Ordinal Regression Network for Monocular Depth Estimationcited by 721
38 more
  1. 2021DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Camerascited by 673
  2. 2016Direct Sparse Odometrycited by 479
  3. 2016Past, Present, and Future of Simultaneous Localization And Mapping: Towards the Robust-Perception Agecited by 405
  4. 2016Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescuecited by 368
  5. 2018GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Posecited by 341
  6. 2019Unsupervised Scale-consistent Depth and Ego-motion Learning from Monocular Videocited by 244
  7. 2018Competitive Collaboration: Joint Unsupervised Learning of Depth, Camera Motion, Optical Flow and Motion Segmentationcited by 222
  8. 2020Consistent Video Depth Estimationcited by 220
  9. 2020D3VO: Deep Depth, Deep Pose and Deep Uncertainty for Monocular Visual Odometrycited by 206
  10. 2016DeMoN: Depth and Motion Network for Learning Monocular Stereocited by 200
  11. 2020Robust Consistent Video Depth Estimationcited by 188
  12. 2018DeepV2D: Video to Depth with Differentiable Structure from Motioncited by 185
  13. 2018Unsupervised Learning of Depth and Ego-Motion from Monocular Video Using 3D Geometric Constraintscited by 179
  14. 2019Depth from Videos in the Wild: Unsupervised Monocular Depth Learning from Unknown Camerascited by 170
  15. 2015Learning Depth from Single Monocular Images Using Deep Convolutional Neural Fieldscited by 169
  16. 2018Depth Prediction Without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videoscited by 155
  17. 2017Learning Depth from Monocular Videos using Direct Methodscited by 154
  18. 2017CNN-SLAM: Real-time dense monocular SLAM with learned depth predictioncited by 152
  19. 2018CodeSLAM - Learning a Compact, Optimisable Representation for Dense Visual SLAMcited by 151
  20. 2018DF-Net: Unsupervised Joint Learning of Depth and Flow using Cross-Task Consistencycited by 148
  21. 2020Feature-metric Loss for Self-supervised Learning of Depth and Egomotioncited by 140
  22. 2019Self-Supervised Monocular Depth Hintscited by 127
  23. 2019Learning the Depths of Moving People by Watching Frozen Peoplecited by 120
  24. 2020DeepFactors: Real-Time Probabilistic Dense Monocular SLAMcited by 117
  25. 2017Semi-Supervised Deep Learning for Monocular Depth Map Predictioncited by 109
  26. 2018Every Pixel Counts ++: Joint Learning of Geometry and Motion with 3D Holistic Understandingcited by 101
  27. 2017UnDeepVO: Monocular Visual Odometry through Unsupervised Deep Learningcited by 95
  28. 2018DeepTAM: Deep Tracking and Mappingcited by 87
  29. 2018Deep Virtual Stereo Odometry: Leveraging Deep Depth Prediction for Monocular Direct Sparse Odometrycited by 85
  30. 2020Towards Better Generalization: Joint Depth-Pose Learning without PoseNetcited by 85
  31. 2018Learning Monocular Depth by Distilling Cross-domain Stereo Networkscited by 64
  32. 2018SuperDepth: Self-Supervised, Super-Resolved Monocular Depth Estimationcited by 64
  33. 2019Learning monocular depth estimation infusing traditional stereo knowledgecited by 64
  34. 2021Consistent Depth of Moving Objects in Videocited by 56
  35. 2019Recurrent Neural Network for (Un-)supervised Learning of Monocular VideoVisual Odometry and Depthcited by 50
  36. 2021Multi-View Depth Estimation by Fusing Single-View Depth Probability with Multi-View Geometrycited by 46
  37. 2020Pseudo RGB-D for Self-Improving Monocular SLAM and Depth Predictioncited by 41
  38. 2017Unsupervised Learning of Geometry with Edge-aware Depth-Normal Consistencycited by 40

50 most cited of 61 All 61 (JSON)

Built on by

Cited by year

Citing papers per year: 2022: 4; 2023: 3; 2024: 0; 2025: 1.

8 papers.

  1. 2022RIAV-MVS: Recurrent-Indexing an Asymmetric Volume for Multi-View Stereocited by 12
  2. 2023SimpleMapping: Real-Time Visual-Inertial Dense Mapping with Deep Multi-View Stereocited by 4
  3. 2022FisheyeDistill: Self-Supervised Monocular Depth Estimation with Ordinal Distillation for Fisheye Camerascited by 3
  4. 2023Dynamic Voxel Grid Optimization for High-Fidelity RGB-D Supervised Surface Reconstructioncited by 3
  5. 2022Semantics-Depth-Symbiosis: Deeply Coupled Semi-Supervised Learning of Semantics and Depthcited by 2
  6. 2023Depth Completion with Multiple Balanced Bases and Confidence for Dense Monocular SLAMcited by 2
  7. 2025An Online Adaptation Method for Robust Depth Estimation and Visual Odometry in the Open Worldcited by 1
  8. 2022CNN-Augmented Visual-Inertial SLAM with Planar Constraintscited by 0