Exploiting uncertainty in regression forests for accurate camera relocalization
The DOI matches its Crossref record; 23 of the 23 references it cites were found in a registry or an index.
Abstract
Recent advances in camera relocalization use predictions from a regression forest to guide the camera pose optimization procedure. In these methods, each tree associates one pixel with a point in the scene's 3D world coordinate frame. In previous work, these predictions were point estimates and the subsequent camera pose optimization implicitly assumed an isotropic distribution of these estimates. In this paper, we train a regression forest to predict mixtures of anisotropic 3D Gaussians and show how the predicted uncertainties can be taken into account for continuous pose optimization. Experiments show that our proposed method is able to relocalize up to 40% more frames than the state of the art.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Likewise it is possible to localize the 6DoF of a camera with regression forest [247]”
From Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016) · cites this paper as [247]
Citations
Cites
Mean shift: a robust approach toward feature space analysis
Comaniciu, D., & Meer, P. (2002) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/34.1000236 · Cited by 11,305 · cited 1×
“Similar to [19], we run mean shift mode detection [3] to find the modes of the empirical distribution in each leaf.”
MethodKinectFusion: Real-time dense surface mapping and tracking
Newcombe, R. A., Fitzgibbon, A., Izadi, S., Hilliges, O., Molyneaux, D., Kim, D., Davison, A. J., Kohi, P., Shotton, J., & Hodges, S. (2011) · 2011 10th IEEE International Symposium on Mixed and Augmented Reality · doi:10.1109/ismar.2011.6092378 · Cited by 4,008 · cited 1×
“Simultaneous Localization and Mapping (SLAM) systems such as [14] [15] [17] [7] and commercial systems such as HoloLens or Project Tango demonstrate that visual SLAM is a maturing technology.”
BackgroundSelf-citationLSD-SLAM: Large-scale direct monocular SLAM
Engel, J., Schöps, T., & Cremers, D. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10605-2_54 · Cited by 3,851 · cited 1×
“Simultaneous Localization and Mapping (SLAM) systems such as [14] [15] [17] [7] and commercial systems such as HoloLens or Project Tango demonstrate that visual SLAM is a maturing technology.”
BackgroundA solution for the best rotation to relate two sets of vectors
Kabsch, W. (1976) · Acta Crystallographica Section A · doi:10.1107/s0567739476001873 · Cited by 3,187 · cited 1×
“When per-pixel depth information is available, one can estimate the transformation parameters from the current view to the model with three local positions (current camera space) and their corresponding positions in the scene (world coordinates) using the Kabsch algorithm [11].”
MethodDTAM: Dense tracking and mapping in real-time
Newcombe, R. A., Lovegrove, S. J., & Davison, A. J. (2011) · 2011 International Conference on Computer Vision · doi:10.1109/iccv.2011.6126513 · Cited by 2,078 · cited 1×
“Simultaneous Localization and Mapping (SLAM) systems such as [14] [15] [17] [7] and commercial systems such as HoloLens or Project Tango demonstrate that visual SLAM is a maturing technology.”
BackgroundReal-time human pose recognition in parts from single depth images
Shotton, J., Sharp, T., Kipman, A., Fitzgibbon, A., Finocchio, M., Blake, A., Cook, M., & Moore, R. (2013) · Communications of the ACM · doi:10.1145/2398356.2398381 · Cited by 2,044 · cited 1×
“Following [19], we use features based on pairwise pixel relationships [13, 20] as they have proven to be well-suited for the task of camera relocalization.”
MethodSelf-citationFAB-MAP: Probabilistic Localization and Mapping in the Space of Appearance
Cummins, M., & Newman, P. (2008) · The International Journal of Robotics Research · doi:10.1177/0278364908090961 · Cited by 1,472 · cited 1×
“A few successful approaches also exist that do not belong to either of those two classes [5, 16].”
BackgroundReal-time 3D reconstruction at scale using voxel hashing
Nießner, M., Zollhöfer, M., Izadi, S., & Stamminger, M. (2013) · ACM Transactions on Graphics · doi:10.1145/2508363.2508374 · Cited by 1,002 · cited 1×
“Simultaneous Localization and Mapping (SLAM) systems such as [14] [15] [17] [7] and commercial systems such as HoloLens or Project Tango demonstrate that visual SLAM is a maturing technology.”
BackgroundSelf-citationScene Coordinate Regression Forests for Camera Relocalization in RGB-D Images
Shotton, J., Glocker, B., Zach, C., Izadi, S., Criminisi, A., & Fitzgibbon, A. (2013) · 2013 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2013.377 · Cited by 930 · cited 3×
“Perhaps the simplest approach, used in [19], is to fit an isotropic Gaussian to the labels in S.”
BackgroundMethodInfluentialSelf-citationChum, O., Matas, J., & Kittler, J. (2003) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-45243-0_31 · Cited by 843 · cited 3×
“Having computed the set of inliers, [2] suggests performing a local optimization step to refine the hypotheses.”
MethodInfluentialTracking people with twists and exponential maps
Bregler, C., & Malik, J. (2002) · Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231) · doi:10.1109/cvpr.1998.698581 · Cited by 719 · cited 1×
“To be able to continuously optimize the proxy energy function over the 6 degrees of freedom of H , we project H from SE(3) to se(3) and obtain the corresponding 6D twist vector ξ = log(H) (see [1]).”
MethodKeypoint recognition using randomized trees
Lepetit, V., & Fua, P. (2006) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2006.188 · Cited by 685 · cited 1×
“Following [19], we use features based on pairwise pixel relationships [13, 20] as they have proven to be well-suited for the task of camera relocalization.”
MethodDecision Forests for Computer Vision and Medical Image Analysis
(2013) · Advances in Computer Vision and Pattern Recognition · doi:10.1007/978-1-4471-4929-3 · Cited by 595
Self-citationVision-based global localization and mapping for mobile robots
Se, S., Lowe, D. G., & Little, J. J. (2005) · IEEE Transactions on Robotics · doi:10.1109/tro.2004.839228 · Cited by 482 · cited 1×
“Publications presenting on-line keypoint-based relocalization methods include [6] [22] [18] [21].”
MethodImproving the Agility of Keyframe-Based SLAM
Klein, G., & Murray, D. (2008) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-88688-4_59 · Cited by 335
Williams, B., Klein, G., & Reid, I. (2007) · 2007 IEEE 11th International Conference on Computer Vision · doi:10.1109/iccv.2007.4409115 · Cited by 224 · cited 1×
“Publications presenting on-line keypoint-based relocalization methods include [6] [22] [18] [21].”
MethodReal-time RGB-D camera relocalization
Glocker, B., Izadi, S., Shotton, J., & Criminisi, A. (2013) · 2013 IEEE International Symposium on Mixed and Augmented Reality (ISMAR) · doi:10.1109/ismar.2013.6671777 · Cited by 195 · cited 1×
“Successful attempts at solving these challenges have been proposed in [9] [8] [12].”
BackgroundSelf-citationUnified Loop Closing and Recovery for Real Time Monocular SLAM
Eade, E. D., & Drummond, T. W. (2008) · Procedings of the British Machine Vision Conference 2008 · doi:10.5244/c.22.6 · Cited by 153 · cited 1×
“Publications presenting on-line keypoint-based relocalization methods include [6] [22] [18] [21].”
MethodAutomatic Relocalization and Loop Closing for Real-Time Monocular SLAM
Williams, B., Klein, G., & Reid, I. D. (2011) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2011.41 · Cited by 101 · cited 1×
“Publications presenting on-line keypoint-based relocalization methods include [6, 22, 18, 21].”
MethodMulti-output Learning for Camera Relocalization
Guzman-Rivera, A., Kohli, P., Glocker, B., Shotton, J., Sharp, T., Fitzgibbon, A., & Izadi, S. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.146 · Cited by 100 · cited 1×
“We first describe the image features and the tree training objective, followed by our method for modeling uncertainty in the predictions made by the forest.”
MethodSelf-citationNi, K., Kannan, A., Criminisi, A., & Winn, J. M. (2009) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2009.165 · Cited by 66
6D Relocalisation for RGBD Cameras Using Synthetic View Regression
Gee, A. P., & Mayol-Cuevas, W. W. (2012) · British Machine Vision Conference (BMVC) · doi:10.5244/c.26.113 · Cited by 41 · cited 1×
“Successful attempts at solving these challenges have been proposed in [9] [8] [12].”
BackgroundNi, K., Kannan, A., Criminisi, A., & Winn, J. M. (2008) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2008.4587585 · Cited by 34 · cited 1×
“A few successful approaches also exist that do not belong to either of those two classes [5] [16].”
Background
Cited by
Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age
Cadena, C., Carlone, L., Carrillo, H., Latif, Y., Scaramuzza, D., Neira, J., Reid, I., & Leonard, J. J. (2016) · IEEE Transactions on Robotics · doi:10.1109/tro.2016.2624754 · Cited by 3,596
“Likewise, it is possible to localize the 6DoF of a camera with regression forest [247] and with deep convolutional neural network [129], and to estimate the depth of a scene (in effect, the map) from a single view solely as a function of the”
Past, present, and future of simultaneous localization and mapping: Toward the robust-perception age
Cadena, C., Carlone, L., Carrillo, H., Latif, Y., Scaramuzza, D., Neira, J., Reid, I. D., & Leonard, J. J. (2016) · Zurich Open Repository and Archive (University of Zurich) · doi:10.5167/uzh-138892 · Cited by 3,048
Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation
Wang, H., Sridhar, S., Huang, J., Valentin, J., Song, S., & Guibas, L. J. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00275 · Cited by 757
Self-citationBundleFusion: Real-Time Globally Consistent 3D Reconstruction Using On-the-Fly Surface Reintegration
Dai, A., Nießner, M., Zollhöfer, M., Izadi, S., & Theobalt, C. (2017) · ACM Transactions on Graphics · doi:10.1145/3072959.3054739 · Cited by 662
Self-citationDSAC — Differentiable RANSAC for Camera Localization
Brachmann, E., Krull, A., Nowozin, S., Shotton, J., Michel, F., Gumhold, S., & Rother, C. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.267 · Cited by 607
“We demonstrate an end-to-end trainable solution for this problem, building on the scene coordinate regression forest (SCoRF) approach [36, 40, 5].”
MethodSelf-citation
Metadata
- Authors
- Valentin, J., Niebner, M., Shotton, J., Fitzgibbon, A., Izadi, S., & Torr, P.
- Affiliations
- University of Oxford (GB)
- Stanford University (US)
- Microsoft (United States) (US)
- Venue
- 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2015-06
- Pages
- 4400-4408
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W1893935112
- Cited by
- 156
- References
- 23
- Citations by year
- 2026: 4 · 2025: 9 · 2024: 13 · 2023: 7 · 2022: 13 · 2021: 19 · 2020: 19 · 2019: 25 · 2018: 20 · 2017: 15 · 2016: 12
- FWCI
- 27.16
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- Funders
- Leverhulme Trust
- Engineering and Physical Sciences Research Council (EP/I001107/2)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Robotics and Sensor-Based Localization
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Vision and Imaging
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Keywords
- camera relocalization, regression forests, uncertainty estimation, camera pose optimization, visual SLAM
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.