DynamicFusion: Reconstruction and tracking of non-rigid scenes in real-time
The DOI matches its Crossref record; 42 of the 42 references it cites were found in a registry or an index.
Abstract
We present the first dense SLAM system capable of reconstructing non-rigidly deforming scenes in real-time, by fusing together RGBD scans captured from commodity sensors. Our DynamicFusion approach reconstructs scene geometry whilst simultaneously estimating a dense volumetric 6D motion field that warps the estimated geometry into a live frame. Like KinectFusion, our system produces increasingly denoised, detailed, and complete reconstructions as more measurements are fused, and displays the updated model in real time. Because we do not require a template or other prior scene model, the approach is applicable to a wide range of moving objects and scenes.
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
“In the SLAM community, Newcombe et al. [182] have address the non-rigid case for small-scale reconstruction.”
From Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016) · cites this paper as [182]
Citations
Cites
KinectFusion: 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 14×
“As in KinectFusion, volumetric TSDF memory limits geometric extent; but there are many solutions to this in the literature.”
MethodInfluentialSelf-citationA volumetric method for building complex models from range images
Curless, B., & Levoy, M. (1996) · Proceedings of the 23rd annual conference on Computer graphics and interactive techniques · doi:10.1145/237170.237269 · Cited by 3,124 · cited 3×
“The truncation distance τ > 0 and the updated TSDF value is given by the weighted averaging scheme [5], with the weight truncation introduced in [18]:”
BackgroundMethodKinectFusion: real-time 3D reconstruction and interaction using a moving depth camera
Izadi, S., Kim, D., Hilliges, O., Molyneaux, D., Newcombe, R., Kohli, P., Shotton, J., Hodges, S., Freeman, D., Davison, A., et al. (2011) · Proceedings of the 24th annual ACM symposium on User interface software and technology · doi:10.1145/2047196.2047270 · Cited by 2,110 · cited 13×
“As in KinectFusion, volumetric TSDF memory limits geometric extent; but there are many solutions to this in the literature.”
MethodInfluentialSelf-citationReal-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×
“KinectFusion spurred a flurry of follow up research aimed at robustifying the tracking [9, 32] and expanding its spatial mapping capabilities to larger environments [22, 19, 34, 31, 9].”
BackgroundEfficient model-based 3D tracking of hand articulations using Kinect
Oikonomidis, I., Kyriazis, N., & Argyros, A. (2011) · Procedings of the British Machine Vision Conference 2011 · doi:10.5244/c.25.101 · Cited by 874 · cited 1×
“The best of these demonstrate high accuracy, real-time performance capture for tracking faces [16, 3], hands [21, 20], complete bodies [27], or general articulated objects [23, 33].”
BackgroundKintinuous: Spatially Extended KinectFusion
Whelan, T. J., Kaess, M., Fallon, M. F., Johannsson, H., Leonard, J. J., & McDonald, J. (2012) · Maynooth University ePrints and eTheses Archive (Maynooth University) · Cited by 624
Realtime and Robust Hand Tracking from Depth
Qian, C., Sun, X., Wei, Y., Tang, X., & Sun, J. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.145 · Cited by 470
Embedded deformation for shape manipulation
Sumner, R. W., Schmid, J., & Pauly, M. (2007) · ACM SIGGRAPH 2007 papers · doi:10.1145/1275808.1276478 · Cited by 449 · cited 3×
“Hierarchical Deformation Tree: In original applications of the embedded deformation graph [25] approach to non-rigid tracking, E is defined as the k−nearest neighbours of each node or all nodes within a specified radius.”
BackgroundInfluentialReal-time non-rigid reconstruction using an RGB-D camera
Zollhöfer, M., Nießner, M., Izadi, S., Rehmann, C., Zach, C., Fisher, M., Wu, C., Fitzgibbon, A., Loop, C., Theobalt, C., et al. (2014) · ACM Transactions on Graphics · doi:10.1145/2601097.2601165 · Cited by 405 · cited 1×
“Recently, [37] demonstrated an impressive real-time version of a similar technique, using GPU accelerated optimisations.”
MethodGlobal Correspondence Optimization for Non‐Rigid Registration of Depth Scans
Li, H., Sumner, R. W., & Pauly, M. (2008) · Computer Graphics Forum · doi:10.1111/j.1467-8659.2008.01282.x · Cited by 399 · cited 1×
“Practical advancements to pairwise 3D shape and scan alignment over larger deformations make use of reduced deformable model parametrisations [14, 4].”
Method
Research works printed but not linked: 1
- curless, 1997, New Methodsfor Surface Reconstructionfrom Range Images
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
“[182] have address the nonrigid case for small-scale reconstruction.”
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
Deferred neural rendering: image synthesis using neural textures
Thies, J., Zollhöfer, M., & Nießner, M. (2019) · ACM Transactions on Graphics · doi:10.1145/3306346.3323035 · Cited by 1,356
“A major benefit of this line of work is that the reconstructed 3D content can be visualized by standard rendering techniques, and the approach directly generalizes to 4D capture as long as non-rigid surfaces can be reliably tracked [Dou et al. 2016; Innmann et al. 2016; Newcombe et al. 2015].”
BackgroundNerfies: Deformable Neural Radiance Fields
Park, K., Sinha, U., Barron, J. T., Bouaziz, S., Goldman, D. B., Seitz, S. M., & Martin-Brualla, R. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00581 · Cited by 1,059
Self-citationVNect: real-time 3D human pose estimation with a single RGB camera
Mehta, D., Sridhar, S., Sotnychenko, O., Rhodin, H., Shafiei, M., Seidel, H.-P., Xu, W., Casas, D., & Theobalt, C. (2017) · ACM Transactions on Graphics · doi:10.1145/3072959.3073596 · Cited by 1,113
“2014] and template-free reconstruction [Dou et al. 2016; Innmann et al. 2016; Newcombe et al. 2015; Orts-Escolano et al. 2016] has been demonstrated.”
MethodHyperNeRF: a higher-dimensional representation for topologically varying neural radiance fields
Park, K., Sinha, U., Hedman, P., Barron, J. T., Bouaziz, S., Goldman, D. B., Martin-Brualla, R., & Seitz, S. M. (2021) · ACM Transactions on Graphics · doi:10.1145/3478513.3480487 · Cited by 618
“Deformation-based NeRF variants follow the tradition established by the significant body of research on non-rigid reconstruction [Newcombe et al. 2015], and map observations of the subject onto a template of that subject.”
MethodSelf-citationNeural volumes: learning dynamic renderable volumes from images
Lombardi, S., Simon, T., Saragih, J., Schwartz, G., Lehrmann, A., & Sheikh, Y. (2019) · ACM Transactions on Graphics · doi:10.1145/3306346.3323020 · Cited by 693
“Over the years, significant efforts have been spent on their reconstruction and tracking from RGB-D sensors: the DynamicFusion method of Newcombe et al. [2015] produces a base 3D template surface using a Truncated Signed Distance Function (TSDF) representation, and a time-dependent warp to fuse a…”
MethodNeural Scene Flow Fields for Space-Time View Synthesis of Dynamic Scenes
Li, Z., Niklaus, S., Snavely, N., & Wang, O. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00643 · Cited by 674
ElasticFusion: Real-time dense SLAM and light source estimation
Whelan, T., Salas-Moreno, R. F., Glocker, B., Davison, A. J., & Leutenegger, S. (2016) · The International Journal of Robotics Research · doi:10.1177/0278364916669237 · Cited by 698
Dynamic 3D Gaussians: Tracking by Persistent Dynamic View Synthesis
Luiten, J., Kopanas, G., Leibe, B., & Ramanan, D. (2024) · 2024 International Conference on 3D Vision (3DV) · doi:10.1109/3dv62453.2024.00044 · Cited by 469
“Such approaches either rely on the presence of accurate depth cameras [27, 34], assume ground-truth object point clouds as input [31], or are heavily specific to certain domains e.g. reconstructing moving cars in driving scenes [3, 23].”
BackgroundPeng, S., Zhang, Y., Xu, Y., Wang, Q., Shuai, Q., Bao, H., & Zhou, X. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00894 · Cited by 626
“Most methods [42, 8, 14, 21] adopt the traditional modeling and rendering pipeline to synthesize novel views of humans.”
MethodA Survey of Surface Reconstruction from Point Clouds
Berger, M., Tagliasacchi, A., Seversky, L. M., Alliez, P., Guennebaud, G., Levine, J. A., Sharf, A., & Silva, C. T. (2017) · Computer Graphics Forum · doi:10.1111/cgf.12802 · Cited by 509
“Recent work [NFS15] has focused on real-time reconstruction of non-rigid scenes by updating a TSDF through graph-based deformations of the captured geometry.”
BackgroundImplicit Functions in Feature Space for 3D Shape Reconstruction and Completion
Chibane, J., Alldieck, T., & Pons-Moll, G. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00700 · Cited by 509
Fusion4D: real-time performance capture of challenging scenes
Dou, M., Khamis, S., Degtyarev, Y., Davidson, P., Fanello, S. R., Kowdle, A., Escolano, S. O., Rhemann, C., Kim, D., Taylor, J., et al. (2016) · ACM Transactions on Graphics · doi:10.1145/2897824.2925969 · Cited by 500
“Recent work has demonstrated compelling real-time reconstructions of general nonrigid scenes using a single depth camera [Zollhöfer et al. 2014; Newcombe et al. 2015].”
BackgroundMethodInfluentialClothCap: seamless 4D clothing capture and retargeting
Pons-Moll, G., Pujades, S., Hu, S., & Black, M. J. (2017) · ACM Transactions on Graphics · doi:10.1145/3072959.3073711 · Cited by 495
“Other works also use a free-form surface (Collet et al. 2015; Innmann et al. 2016; Newcombe et al. 2015; Wang et al. 2016) but can not separate the clothing from the person and can not retarget it.”
BackgroundVideo Based Reconstruction of 3D People Models
Alldieck, T., Magnor, M., Xu, W., Theobalt, C., & Pons-Moll, G. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00875 · Cited by 468
Tretschk, E., Tewari, A., Golyanik, V., Zollhofer, M., Lassner, C., & Theobalt, C. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01272 · Cited by 432
“Such sensors can been used for 3D reconstruction and completion of rigid environments [51] and non-rigid objects [98, 50, 28, 71].”
MethodTewari, A., Thies, J., Mildenhall, B., Srinivasan, P., Tretschk, E., Yifan, W., Lassner, C., Sitzmann, V., Martin‐Brualla, R., Lombardi, S., et al. (2022) · Computer Graphics Forum · doi:10.1111/cgf.14507 · Cited by 339
“These SDF representations are used in numerous 3D scanning techniques that use volumetric fusion [CL96] to incrementally reconstruct the surface of a static [IKH ∗ 11, NZIS13] or dynamic object [NFS15].”
MethodMaskFusion: Real-Time Recognition, Tracking and Reconstruction of Multiple Moving Objects
Runz, M., Buffier, M., & Agapito, L. (2018) · 2018 IEEE International Symposium on Mixed and Augmented Reality (ISMAR) · doi:10.1109/ismar.2018.00024 · Cited by 415
Multi-Garment Net: Learning to Dress 3D People From Images
Bhatnagar, B., Tiwari, G., Theobalt, C., & Pons-Moll, G. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00552 · Cited by 439
“Without a pre-scanned template, fusion [30, 29, 55, 43, 58] trackers incrementally fuse geometry and appearance [66] to build the template on the fly.”
BackgroundAnimatable Neural Radiance Fields for Modeling Dynamic Human Bodies
Peng, S., Dong, J., Wang, Q., Zhang, S., Shuai, Q., Zhou, X., & Bao, H. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01405 · Cited by 387
Space-time Neural Irradiance Fields for Free-Viewpoint Video
Xian, W., Huang, J.-B., Kopf, J., & Kim, C. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00930 · Cited by 357
“Examples include 3D reconstruction for static scenes [49, 27] or dynamic objects from a single [45, 26] or RGB-D cameras [14, 5, 26, 45, 73, 79].”
BackgroundLearning to Dress 3D People in Generative Clothing
Ma, Q., Yang, J., Ranjan, A., Pujades, S., Pons-Moll, G., Tang, S., & Black, M. J. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00650 · Cited by 364
“rgeting it to new body shapes. Similar to ClothCap, Alldieck et al. [35,36] represent clothing as an offset from the SMPL to reconstruct people in clothing from images. Combining dynamic fusion ideas [37] with SMPL, cloth capture and retargeting has been demonstrated from a single depth camera [38,39]. Still other wor…”
Function4D: Real-time Human Volumetric Capture from Very Sparse Consumer RGBD Sensors
Yu, T., Zheng, Z., Guo, K., Liu, P., Dai, Q., & Liu, Y. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00569 · Cited by 355
“3(e)) as in DynamicFu-sion [33] to make sure that all of the surfaces can be covered by the final node graph (Fig.”
BackgroundMethodInfluentialA Survey on Active Simultaneous Localization and Mapping: State of the Art and New Frontiers
Placed, J. A., Strader, J., Carrillo, H., Atanasov, N., Indelman, V., Carlone, L., & Castellanos, J. A. (2023) · IEEE Transactions on Robotics · doi:10.1109/tro.2023.3248510 · Cited by 332
“Medical applications have also stimulated progress on SLAM in deformable environments [263], [264].”
BackgroundDynamic FAUST: Registering Human Bodies in Motion
Bogo, F., Romero, J., Pons-Moll, G., & Black, M. J. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.591 · Cited by 358
“For example, there are many non-rigid tracking methods, either model-based [3, 7, 22, 25, 27, 34, 44, 49] or model-free [2, 16, 21, 24, 37, 50, 54].”
BackgroundDeepHuman: 3D Human Reconstruction From a Single Image
Zheng, Z., Yu, T., Wei, Y., Dai, Q., & Liu, Y. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00783 · Cited by 368
ARCH: Animatable Reconstruction of Clothed Humans
Huang, Z., Xu, Y., Lassner, C., Li, H., & Tung, T. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00316 · Cited by 341
Huang, Y., Kaufmann, M., Aksan, E., Black, M. J., Hilliges, O., & Pons-Moll, G. (2018) · ACM Transactions on Graphics · doi:10.1145/3272127.3275108 · Cited by 342
Learning to Reconstruct People in Clothing From a Single RGB Camera
Alldieck, T., Magnor, M., Bhatnagar, B. L., Theobalt, C., & Pons-Moll, G. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00127 · Cited by 326
“DynamicFusion [51] generalized KinectFusion to non-rigid objects by combining non-rigid tracking and fusion.”
BackgroundKimera: From SLAM to spatial perception with 3D dynamic scene graphs
Rosinol, A., Violette, A., Abate, M., Hughes, N., Chang, Y., Shi, J., Gupta, A., & Carlone, L. (2021) · The International Journal of Robotics Research · doi:10.1177/02783649211056674 · Cited by 271
“Approaches that deal with moving objects, such as DynamicFusion (Newcombe et al. 2015), Mask-fusion (Runz et al. 2018), Co-fusion (Rünz and Agapito 2017), and MID-Fusion (Xu et al. 2019) are currently limited to small table-top scenes and focus on objects or dense maps, rather than scene graphs.”
BackgroundVolumeDeform: Real-Time Volumetric Non-rigid Reconstruction
Innmann, M., Zollhöfer, M., Nießner, M., Theobalt, C., & Stamminger, M. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46484-8_22 · Cited by 308
“The depth data D t of each recorded RGB-D frame is incrementally fused into the canonical TSDF following the non-rigid fusion technique introduced in DynamicFusion [17].”
BackgroundMethodInfluentialDetailed, Accurate, Human Shape Estimation from Clothed 3D Scan Sequences
Zhang, C., Pujades, S., Black, M., & Pons-Moll, G. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.582 · Cited by 295
“Other work is purely model free and attempts to estimate non-rigid 3D shape over time [41, 22, 11, 25].”
BackgroundModeling, learning, perception, and control methods for deformable object manipulation
Yin, H., Varava, A., & Kragic, D. (2021) · Science Robotics · doi:10.1126/scirobotics.abd8803 · Cited by 299
“If a canonical template is not available, it can be iteratively updated (55).”
BackgroundSC-GS: Sparse-Controlled Gaussian Splatting for Editable Dynamic Scenes
Huang, Y.-H., Sun, Y.-T., Yang, Z., Lyu, X., Cao, Y.-P., & Qi, X. (2024) · 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52733.2024.00404 · Cited by 183
“Then, the interpolation weights for control point p k can be com-3 puted with Gaussian-kernel RBF [4, 9, 32] as: where d jk is the distance between center of Gaussian G j and the neighboring control point p k , and o k is the learned radius parameter of p k .”
BackgroundMethodYu, T., Zheng, Z., Guo, K., Zhao, J., Dai, Q., Li, H., Pons-Moll, G., & Liu, Y. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00761 · Cited by 255
3D Dynamic Scene Graphs: Actionable Spatial Perception with Places, Objects, and Humans
Rosinol, A., Gupta, A., Abate, M., Shi, J., & Carlone, L. (2020) · Robotics: Science and Systems XVI · doi:10.15607/rss.2020.xvi.079 · Cited by 201
A review of visual SLAM methods for autonomous driving vehicles
Cheng, J., Zhang, L., Chen, Q., Hu, X., & Cai, J. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.104992 · Cited by 252
SE3-nets: Learning rigid body motion using deep neural networks
Byravan, A., & Fox, D. (2017) · 2017 IEEE International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra.2017.7989023 · Cited by 236
“Another exciting direction is to use the “No Penalty” version of SE3-NETS to model nonrigid motion, exploring the use of strong regularization and locality priors in the masks [33] to improve efficiency and”
MethodSelf-citationThe relightables: volumetric performance capture of humans with realistic relighting
Guo, K., Lincoln, P., Davidson, P., Busch, J., Yu, X., Whalen, M., Harvey, G., Orts-Escolano, S., Pandey, R., Dourgarian, J., et al. (2019) · ACM Transactions on Graphics · doi:10.1145/3355089.3356571 · Cited by 236
“An alignment algorithm [Newcombe et al. 2015] consumes the sequence of reconstructed meshes so that long subsequences can ACM Transactions on Graphics, Vol. 38, No. 6, Article 217.”
BackgroundMethodLiveCap: Real-Time Human Performance Capture From Monocular Video
Habermann, M., Xu, W., Zollhöfer, M., Pons-Moll, G., & Theobalt, C. (2019) · ACM Transactions on Graphics · doi:10.1145/3311970 · Cited by 230
“Although Although these methods are a significant step forward, RGB-D cameras are not as cheap and ubiquitous as color cameras, often have a limited capture range, do not work well under bright sunlight, and have limited resolution.”
BackgroundMethodInfluentialMonoPerfCap: Human Performance Capture From Monocular Video
Xu, W., Chatterjee, A., Zollhöfer, M., Rhodin, H., Mehta, D., Seidel, H.-P., & Theobalt, C. (2018) · ACM Transactions on Graphics · doi:10.1145/3181973 · Cited by 215
“2015;Shotton et al. 2011] demonstrate the possibility of low cost performance capture with commodity hardware. Even real-time tracking [Zollhöfer et al. 2014] and reconstruction [Innmann et al. 2016;Newcombe et al. 2015] of general deforming objects has been demonstrated. While reliable in controlled indoor settings,…”
InfluentialCvxNet: Learnable Convex Decomposition
Deng, B., Genova, K., Yazdani, S., Bouaziz, S., Hinton, G., & Tagliasacchi, A. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00011 · Cited by 215
“In computer vision and robotics, analogous trade-offs exist: surface models are essential for the construction of low-dimensional parametric templates essential for tracking [6, 8], volumetric representations are key to capturing 3D data whose topology is unknown [48, 47], while part-based models provide a natural dec…”
BackgroundDeformable Shape Completion with Graph Convolutional Autoencoders
Litany, O., Bronstein, A., Bronstein, M., & Makadia, A. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00202 · Cited by 226
Near‐Eye Display and Tracking Technologies for Virtual and Augmented Reality
Koulieris, G. A., Akşit, K., Stengel, M., Mantiuk, R. K., Mania, K., & Richardt, C. (2019) · Computer Graphics Forum · doi:10.1111/cgf.13654 · Cited by 220
Learning the Depths of Moving People by Watching Frozen People
Li, Z., Dekel, T., Cole, F., Tucker, R., Snavely, N., Liu, C., & Freeman, W. T. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00465 · Cited by 224
“RGBD data has been widely used for 3D modeling of dynamic scenes [25, 55, 48, 5, 14], but only a few methods attempt to estimate depth from a monocular camera.”
MethodCo-fusion: Real-time segmentation, tracking and fusion of multiple objects
Runz, M., & Agapito, L. (2017) · 2017 IEEE International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra.2017.7989518 · Cited by 227
“DynamicFusion [13] is a prime example of a monocular real time system that can fuse together scans of deformable objects captured from depth sensors without the need for any pre-trained model or shape template.”
BackgroundDeepWrinkles: Accurate and Realistic Clothing Modeling
Lähner, Z., Cremers, D., & Tung, T. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-01225-0_41 · Cited by 225
MID-Fusion: Octree-based Object-Level Multi-Instance Dynamic SLAM
Xu, B., Li, W., Tzoumanikas, D., Bloesch, M., Davison, A., & Leutenegger, S. (2019) · 2019 International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra.2019.8794371 · Cited by 217
“One deforms the whole world in a non-rigid manner in order to include a deformable/moving object [7].”
BackgroundDense Object Nets: Learning Dense Visual Object Descriptors By and For Robotic Manipulation
Florence, P., Manuelli, L., & Tedrake, R. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1806.08756 · Cited by 142
“Prior work [7] has used dynamic reconstruction [21] of raw RGBD data for only within-scene data association and remarkably showed that even without cross-scene data association, descriptors could be learned that were consistent across many dynamic scenes of the upper body of a human subject.”
MethodDeepCap: Monocular Human Performance Capture Using Weak Supervision
Habermann, M., Xu, W., Zollhofer, M., Pons-Moll, G., & Theobalt, C. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00510 · Cited by 214
Detailed Human Avatars from Monocular Video
Alldieck, T., Magnor, M., Xu, W., Theobalt, C., & Pons-Moll, G. (2018) · 2018 International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2018.00022 · Cited by 206
Lepard: Learning partial point cloud matching in rigid and deformable scenes
Li, Y., & Harada, T. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00547 · Cited by 186
“DynamicFusion [46] employs the simple projective correspondence for real-time efficiency.”
BackgroundTensor4D: Efficient Neural 4D Decomposition for High-Fidelity Dynamic Reconstruction and Rendering
Shao, R., Zheng, Z., Tu, H., Liu, B., Zhang, H., & Liu, Y. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.01596 · Cited by 180
“With RGBD cameras, real-time solutions like DynamicFusion [34] estimates the non-rigid deformations of a dynamic scene and integrates depth frames to reconstruct the geometry model in the canonical space.”
MethodDetailed Full-Body Reconstructions of Moving People from Monocular RGB-D Sequences
Bogo, F., Black, M. J., Loper, M., & Romero, J. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.265 · Cited by 200
“Recently, [25] extends KinectFusion to capture dynamic 3D shapes including partial views of moving people.”
BackgroundCombining Implicit Function Learning and Parametric Models for 3D Human Reconstruction
Bhatnagar, B. L., Sminchisescu, C., Theobalt, C., & Pons-Moll, G. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58536-5_19 · Cited by 200
Recasens, D., Lamarca, J., Facil, J. M., Montiel, J. M. M., & Civera, J. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3095528 · Cited by 175
DynIBaR: Neural Dynamic Image-Based Rendering
Li, Z., Wang, Q., Cole, F., Tucker, R., & Snavely, N. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.00416 · Cited by 174
SiCloPe: Silhouette-Based Clothed People
Natsume, R., Saito, S., Huang, Z., Chen, W., Ma, C., Li, H., & Morishima, S. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00461 · Cited by 199
“While 3D range sensing [24, 32] and photogramme-try [37] are popular ways of obtaining complete 3D models, they are restricted to a tedious scanning process or require specialized equipment.”
BackgroundDense Human Body Correspondences Using Convolutional Networks
Wei, L., Huang, Q., Ceylan, D., Vouga, E., & Li, H. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.171 · Cited by 185
“With the popularization of commodity 3D scanners and recent advances in correspondence algorithms for deformable shapes, human bodies can now be easily digitized [24, 29, 11] and their performances captured using a single RGB-D sensor [23, 42].”
BackgroundiSDF: Real-Time Neural Signed Distance Fields for Robot Perception
Ortiz, J., Clegg, A., Dong, J., Sucar, E., Novotný, D., Zollhöfer, M., & Mukadam, M. (2022) · Robotics: Science and Systems · doi:10.15607/rss.2022.xviii.012 · Cited by 237
BundleSDF: Neural 6-DoF Tracking and 3D Reconstruction of Unknown Objects
Wen, B., Tremblay, J., Blukis, V., Tyree, S., Müller, T., Evans, A., Fox, D., Kautz, J., & Birchfield, S. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.00066 · Cited by 166
“Note that methods such as [20, 42] focus on deformable objects and the root 6-DoF tracking and fusion are often based on [43], whereas we focus on rigid objects that are dynamically moving.”
BackgroundMethodSelf-citationShape of Motion: 4D Reconstruction From a Single Video
Wang, Q., Ye, V., Gao, H., Zeng, W., Austin, J., Li, Z., & Kanazawa, A. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00901 · Cited by 287
“In non-rigid reconstruction, early methods often required RGB-D sensors [5, 19, 41, 73, 130] or strong hand-crafted priors [54, 81, 85].”
KillingFusion: Non-rigid 3D Reconstruction without Correspondences
Slavcheva, M., Baust, M., Cremers, D., & Ilic, S. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.581 · Cited by 177
“DynamicFusion [32] was the first approach to simultaneously reconstruct and track the surface motion in real time.”
BackgroundMethodInfluentialBodyFusion: Real-Time Capture of Human Motion and Surface Geometry Using a Single Depth Camera
Yu, T., Guo, K., Xu, F., Dong, Y., Su, Z., Zhao, J., Li, J., Dai, Q., & Liu, Y. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.104 · Cited by 177
“For the first frame, following [27], we rigidly integrate the depth information into the reference volume, extract a triangular surface from the reference volume using the marching cube algorithm [24], uniformly sample deformation nodes on the surface and construct a node graph to describe the non-rigid deformation.”
BackgroundMethodInfluentialNeural Rendering for Stereo 3D Reconstruction of Deformable Tissues in Robotic Surgery
Wang, Y., Long, Y., Fan, S. H., & Dou, Q. (2022) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-16449-1_41 · Cited by 166
Motion2fusion: real-time volumetric performance capture
Dou, M., Davidson, P., Fanello, S. R., Khamis, S., Kowdle, A., Rhemann, C., Tankovich, V., & Izadi, S. (2017) · ACM Transactions on Graphics · doi:10.1145/3130800.3130801 · Cited by 165
“Unlike the dual-quaternion parameterization in DynamicFusion [Newcombe et al. 2015], each ED node “rotates” vertex v around its location g , which makes the optimization problem numerically more stable.”
InfluentialRotation-Invariant Transformer for Point Cloud Matching
Yu, H., Qin, Z., Hou, J., Saleh, M., Li, D., Busam, B., & Ilić, S. (2023) · 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/cvpr52729.2023.00521 · Cited by 248
“The correspondence estimation between a pair of partially-overlapping point clouds is a long-standing task that lies at the core of many computer vision applications, such as tracking [17, 18], reconstruction [22, 23, 39], pose estimation [19, 27, 50] and 3D representation learning [15, 16, 46], etc.”
BackgroundSphere-meshes for real-time hand modeling and tracking
Tkach, A., Pauly, M., & Tagliasacchi, A. (2016) · ACM Transactions on Graphics · doi:10.1145/2980179.2980226 · Cited by 169
“Dynamic reconstruction frameworks such as [Newcombe et al. 2015] or [Innmann et al.”
BackgroundMethodInfluentialA Comprehensive Review of Vision-Based 3D Reconstruction Methods
Zhou, L., Wu, G., Zuo, Y., Chen, X., & Hu, H. (2024) · Sensors · doi:10.3390/s24072314 · Cited by 190
“Spherical harmonics can attenuate high frequencies to a certain extent.”
BackgroundHumanNeRF: Efficiently Generated Human Radiance Field from Sparse Inputs
Zhao, F., Yang, W. H., Zhang, J., Lin, P., Zhang, Y., Yu, J., & Xu, L. (2022) · 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/cvpr52688.2022.00759 · Cited by 208
“Other monocular RGB-D based methods [12, 19, 30, 43, 57, 63] adopt the traditional modeling and rendering pipeline to synthesize novel views of humans.”
MethodGao, W., & Tedrake, R. (2019) · 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.2019.01135 · Cited by 183
“The proposed method is compared with the projective ICP, the original non-rigid tracker of [26].”
BackgroundMethodInfluentialDeep Fashion3D: A Dataset and Benchmark for 3D Garment Reconstruction from Single Images
Zhu, H., Cao, Y., Jin, H., Chen, W., Du, D., Wang, Z., Cui, S. R., & Han, X. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58452-8_30 · Cited by 213
Self-Supervised Visual Descriptor Learning for Dense Correspondence
Schmidt, T., Newcombe, R., & Fox, D. (2017) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2016.2634089 · Cited by 148
“KinectFusion ([17]), for rigid scenes, or DynamicFusion ([18]), for non-rigidly deforming scenes.”
BackgroundMethodSelf-citationMoSca: Dynamic Gaussian Fusion from Casual Videos via 4D Motion Scaffolds
Lei, J., Weng, Y., Harley, A. W., Guibas, L. J., & Daniilidis, K. (2025) · 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/cvpr52734.2025.00578 · Cited by 228
“A few recent works [51, 58, 119, 120] fuse information across frames, but only from a small temporal window.”
Real-Time Geometry, Albedo, and Motion Reconstruction Using a Single RGB-D Camera
Guo, K., Xu, F., Yu, T., Liu, X., Dai, Q., & Liu, Y. (2017) · ACM Transactions on Graphics · doi:10.1145/3072959.3083722 · Cited by 151
“We follow DynamicFusion [23] in representing the motion field W t by a graph-based motion representation, which can effectively represent non-rigid deformation for a surface of any shape and can be applied to deform voxels.”
BackgroundMethodResultInfluentialAdaptive 3D Face Reconstruction from Unconstrained Photo Collections
Roth, J., Tong, Y., & Liu, X. (2016) · 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.2016.455 · Cited by 143
Fourier PlenOctrees for Dynamic Radiance Field Rendering in Real-time
Wang, L., Zhang, J., Liu, X., Zhao, F., Zhang, Y., Zhang, Y., Wu, M., Jingyi Yu, A., & Xu, L. (2022) · 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/cvpr52688.2022.01316 · Cited by 211
Vid2Avatar: 3D Avatar Reconstruction from Videos in the Wild via Self-supervised Scene Decomposition
Guo, C., Jiang, T., Chen, X., Song, J., & Hilliges, O. (2023) · 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/cvpr52729.2023.01236 · Cited by 213
“Depth-based approaches [6, 35, 36] reconstruct the human shape by fusing depth measurements across time.”
BackgroundSensing and Automation in Pruning of Apple Trees: A Review
He, L., & Schupp, J. R. (2018) · Agronomy · doi:10.3390/agronomy8100211 · Cited by 139
“Meanwhile, the serial robot arm with an end-effector requires large space for the cutter to engage with the branches.”
BackgroundRobust Dense Mapping for Large-Scale Dynamic Environments
Bârsan, I. A., Liu, P., Pollefeys, M., & Geiger, A. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8462974 · Cited by 152
“, pedestrians) from multiple observations is more computationally demanding [18].”
BackgroundReal-Time Large-Scale Dense 3D Reconstruction with Loop Closure
Kähler, O., Prisacariu, V. A., & Murray, D. W. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46484-8_30 · Cited by 151
“The prompt delivery of highly detailed dense reconstructions of the 3D environment is currently one of the major frontiers in computer vision [1–6].”
BackgroundSong, J., Wang, J., Zhao, L., Huang, S., & Dissanayake, G. (2018) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2018.2856519 · Cited by 130
SnapToReality: Aligning Augmented Reality to the Real World
Nuernberger, B., Ofek, E., Benko, H., & Wilson, A. D. (2016) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/2858036.2858250 · Cited by 139
“More sophisticated surface reconstruction methods such as Kinect Fusion [25] or DynamicFusion [24] might help.”
MethodRoo, J. S., & Hachet, M. (2017) · ACM Symposium on User Interface Software and Technology (UIST) · doi:10.1145/3126594.3126638 · Cited by 125
“In the same line, full nonrigid reconstruction [31] could be use to scan and track physical elements.”
BackgroundDefSLAM: Tracking and Mapping of Deforming Scenes From Monocular Sequences
Lamarca, J., Parashar, S., Bartoli, A., & Montiel, J. M. M. (2020) · IEEE Transactions on Robotics · doi:10.1109/tro.2020.3020739 · Cited by 125
“DynamicFusion [15] is a seminal work in deformable VSLAM with an RGB-D camera.”
BackgroundSuPer: A Surgical Perception Framework for Endoscopic Tissue Manipulation With Surgical Robotics
Li, Y., Richter, F., Lu, J., Funk, E. K., Orosco, R. K., Zhu, J., & Yip, M. C. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.2970659 · Cited by 155
“The rigidness assumption was then removed to capture the motion of a deforming scene [12].”
BackgroundNeRF-DS: Neural Radiance Fields for Dynamic Specular Objects
Yan, Z., Li, C., & Lee, G. H. (2023) · 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/cvpr52729.2023.00801 · Cited by 123
“A more general approach is to predict a canonical space and a temporal deformation field [7,23,26,28,56].”
BackgroundFast odometry and scene flow from RGB-D cameras based on geometric clustering
Jaimez, M., Kerl, C., González-Jiménez, J., & Cremers, D. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989459 · Cited by 132
“DynamicFusion [21] estimates the camera motion using KinectFusion [8], which is not prepared to handle moving objects.”
MethodLookinGood: enhancing performance capture with real-time neural re-rendering
Martin-Brualla, R., Pandey, R., Yang, S., Pidlypenskyi, P., Taylor, J., Valentin, J., Khamis, S., Davidson, P., Tkach, A., Lincoln, P., et al. (2018) · ACM Transactions on Graphics · doi:10.1145/3272127.3275099 · Cited by 106
Self-citationClusterVO: Clustering Moving Instances and Estimating Visual Odometry for Self and Surroundings
Huang, J., Yang, S. Y., Mu, T., & Hu, S. (2020) · 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/cvpr42600.2020.00224 · Cited by 133
Rate-Utility Optimized Streaming of Volumetric Media for Augmented Reality
Park, J., Chou, P. A., & Hwang, J. (2019) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2019.2898622 · Cited by 120
Self-Supervised Correspondence in Visuomotor Policy Learning
Florence, P., Manuelli, L., & Tedrake, R. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2019.2956365 · Cited by 110
“Other prior work [6] has used dynamic non-rigid reconstruction [35] to address dynamic scenes.”
Learning High Fidelity Depths of Dressed Humans by Watching Social Media Dance Videos
Jafarian, Y., & Park, H. S. (2021) · 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/cvpr46437.2021.01256 · Cited by 237
“1An affine transformation used in the non-rigid tracking [37] can be a complementary to the SE3.”
MethodEditable free-viewpoint video using a layered neural representation
Zhang, J., Liu, X., Ye, X., Zhao, F., Zhang, Y., Wu, M., Zhang, Y., Xu, L., & Yu, J. (2021) · ACM Transactions on Graphics · doi:10.1145/3450626.3459756 · Cited by 177
Label Fusion: A Pipeline for Generating Ground Truth Labels for Real RGBD Data of Cluttered Scenes
Marion, P., Florence, P., Manuelli, L., & Tedrake, R. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8460950 · Cited by 125
“surfel cloud representation in ElasticFusion, and (ii) labeling deformable and dynamic objects using dynamic reconstruction techniques such as [30].”
MethodHigh-quality indoor scene 3D reconstruction with RGB-D cameras: A brief review
Li, J., Gao, W., Wu, Y., Liu, Y., & Shen, Y. (2022) · Computational Visual Media · doi:10.1007/s41095-021-0250-8 · Cited by 124
“, DynamicFusion [18] and DeepHuman [119]).”
BackgroundInfluential4DComplete: Non-Rigid Motion Estimation Beyond the Observable Surface
Li, Y., Takehara, H., Taketomi, T., Zheng, B., & Nießner, M. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01247 · Cited by 166
“ARAP [47] is widely used as a deformation prior in non-rigid reconstruction [38, 26, 59].”
BackgroundMethodInfluentialSobolevFusion: 3D Reconstruction of Scenes Undergoing Free Non-rigid Motion
Slavcheva, M., Baust, M., & Ilić, S. (2018) · 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.2018.00280 · Cited by 123
“Similarly, in Figure 7 we demonstrate better preservation of detail than KillingFusion [40] on the 360◦ Squeeze sequence from DynamicFusion [31].”
BackgroundMethodInfluentialDeFeat-Net: General Monocular Depth via Simultaneous Unsupervised Representation Learning
Spencer, J., Bowden, R., & Hadfield, S. (2020) · 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/cvpr42600.2020.01441 · Cited by 138
Metadata
- Authors
- Newcombe, R. A., Fox, D., & Seitz, S. M.
- Affiliations
- University of Washington (US)
- Venue
- 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2015-06
- Pages
- 343-352
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W1938204631
- Cited by
- 951
- References
- 42
- Citations by year
- 2026: 28 · 2025: 61 · 2024: 67 · 2023: 93 · 2022: 84 · 2021: 113 · 2020: 105 · 2019: 124 · 2018: 110 · 2017: 78 · 2016: 80 · 2015: 8
- FWCI
- 159.57
- 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)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Robotics and Sensor-Based Localization
- Physical Sciences › Earth and Planetary Sciences › Geology › 3D Surveying and Cultural Heritage
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Vision and Imaging
- Keywords
- dense SLAM, real-time 3D reconstruction, KinectFusion
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.