DTAM: Dense tracking and mapping in real-time
The DOI matches its Crossref record; 16 of the 17 references it cites were found in a registry or an index.
Abstract
DTAM is a system for real-time camera tracking and reconstruction which relies not on feature extraction but dense, every pixel methods. As a single hand-held RGB camera flies over a static scene, we estimate detailed textured depth maps at selected keyframes to produce a surface patchwork with millions of vertices. We use the hundreds of images available in a video stream to improve the quality of a simple photometric data term, and minimise a global spatially regularised energy functional in a novel non-convex optimisation framework. Interleaved, we track the camera's 6DOF motion precisely by frame-rate whole image alignment against the entire dense model. Our algorithms are highly parallelisable throughout and DTAM achieves real-time performance using current commodity GPU hardware. We demonstrate that a dense model permits superior tracking performance under rapid motion compared to a state of the art method using features; and also show the additional usefulness of the dense model for real-time scene interaction in a physics-enhanced augmented reality application.
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
“However, in our network, multi-scale projections not only help to overcome gradient locality during training (see [33], [47] for a broader discussion of this well-known issue), but also aid in online error correction at runtime.”
From Unsupervised Deep Visual-Inertial Odometry with Online Error Correction for RGB-D Imagery (2020) · cites this paper as [33]
In the citing paper
“SFMLearner [33] demonstrated the unsupervised learning of unscaled egomotion and depth from RGB imagery.”
From Unsupervised Deep Visual-Inertial Odometry with Online Error Correction for RGB-D Imagery (2020) · cites this paper as [33]
In the citing paper
“For KITTI, VIOLearner RGB outperforms approaches SFMLearner [33] and the network of Zhan et al. [30].”
From Unsupervised Deep Visual-Inertial Odometry with Online Error Correction for RGB-D Imagery (2020) · cites this paper as [33]
In the citing paper
“For AirSim, VIOLearner (no IMU) outperforms approaches SFMLearner [33] and ORB-SLAM2-S in translation error.”
From Unsupervised Deep Visual-Inertial Odometry with Online Error Correction for RGB-D Imagery (2020) · cites this paper as [33]
In the citing paper
“Compared to other vision-only approaches, for KITTI, VIOLearner (no IMU) outperforms approaches SFMLearner [33] and the network of Zhan et al. [30] and yet is outperformed by ORB-SLAM-2 which uses full bundle adjustment.”
From Unsupervised Deep Visual-Inertial Odometry with Online Error Correction for RGB-D Imagery (2020) · cites this paper as [33]
Citations
Cites
Nonlinear total variation based noise removal algorithms
Rudin, L. I., Osher, S., & Fatemi, E. (1992) · Physica D: Nonlinear Phenomena · doi:10.1016/0167-2789(92)90242-f · Cited by 15,704
Parallel Tracking and Mapping for Small AR Workspaces
Klein, G., & Murray, D. (2007) · 2007 6th IEEE and ACM International Symposium on Mixed and Augmented Reality · doi:10.1109/ismar.2007.4538852 · Cited by 4,346
A First-Order Primal-Dual Algorithm for Convex Problems with Applications to Imaging
Chambolle, A., & Pock, T. (2011) · Journal of Mathematical Imaging and Vision · doi:10.1007/s10851-010-0251-1 · Cited by 4,336
Lucas-Kanade 20 Years On: A Unifying Framework
Baker, S., & Matthews, I. (2004) · International Journal of Computer Vision · doi:10.1023/b:visi.0000011205.11775.fd · Cited by 3,155
Fast cost-volume filtering for visual correspondence and beyond
Rhemann, C., Hosni, A., Bleyer, M., Rother, C., & Gelautz, M. (2011) · CVPR 2011 · doi:10.1109/cvpr.2011.5995372 · Cited by 485
Live dense reconstruction with a single moving camera
Newcombe, R. A., & Davison, A. J. (2010) · 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2010.5539794 · Cited by 366
Self-citationImproving 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
Robust multi-sensor image alignment
Irani, M., & Anandan, P. (2002) · Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271) · doi:10.1109/iccv.1998.710832 · Cited by 220
Real-Time Dense Geometry from a Handheld Camera
Stühmer, J., Gumhold, S., & Cremers, D. (2010) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-15986-2_2 · Cited by 205
Some First-Order Algorithms for Total Variation Based Image Restoration
Aujol, J. (2009) · Journal of Mathematical Imaging and Vision · doi:10.1007/s10851-009-0149-y · Cited by 151
Cited by
Items 501–600 of 1,004
Inertial aided dense & semi-dense methods for robust direct visual odometry
Falquez, J. M., Kasper, M. E., & Sibley, G. (2016) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2016.7759530 · Cited by 42
“Recently, many approaches have been developed for direct visual odometry ranging from fully dense methods [12] [14] [7] to semi-dense [2] [13], and even hybrid approaches like semi-direct [3].”
BackgroundNGEL-SLAM: Neural Implicit Representation-based Global Consistent Low-Latency SLAM System
Mao, Y., Yu, X., Zhang, Z., Wang, K., Wang, Y., Xiong, R., & Liao, Y. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10611269 · Cited by 34
“Visual SLAM systems can be categorized into traditional SLAM systems [6], [7], [6], [8], [9], [10], [1], [2], [3] and learning-based SLAM systems [11], [12], [13].”
BackgroundSun, L. C., Bhatt, N. P., Liu, J. C., Fan, Z., Wang, Z., Humphreys, T. E., & Topcu, U. (2024) · 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10802389 · Cited by 34
“Recent advancements in neural rendering, exemplified by NeRFs [15] and subsequent developments [16], [17], have demonstrated significant progress in novel view synthesis using foundational 3D representations such as MLPs, voxel grids, or hash tables.”
Gutiérrez-Gómez, D., Mayol-Cuevas, W. W., & Guerrero, J. J. (2015) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2015.7138984 · Cited by 33
“Thus WA(p) becomes an unknown yielding an ill-posed problem with WimHim constraints for WimHim + 6 unknowns, which is solved using optical flow variational methods [28], or by performing variable baseline stereo matching [7].”
BackgroundMethodEfficient 3D Scene Labeling Using Fields of Trees
Kähler, O., & Reid, I. D. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.380 · Cited by 42
“Remarkable achievements have been made in both of these directions, with systems like PTAM [8], DTAM [9] and KinectFusion [10] able to provide 3D information in real time.”
BackgroundMethodMonocular, Real-Time Surface Reconstruction Using Dynamic Level of Detail
Zienkiewicz, J., Tsiotsios, A., Davison, A. J., & Leutenegger, S. (2016) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2016.82 · Cited by 38
“This approach runs in constant time thanks to the use of orthogonal integral images [34], and it avoids the computational complexity of the global optimisation used in DTAM, while still offering good regularisation properties in low-texture areas and preserving sharp edges.”
BackgroundMethodSelf-citationDivide and Conquer: Efficient Density-Based Tracking of 3D Sensors in Manhattan Worlds
Zhou, Y. B., Kneip, L., Rodriguez, C., & Li, H. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-54193-8_1 · Cited by 34
Single-View and Multi-View Depth Fusion
Facil, J. M., Concha, A., Montesano, L., & Civera, J. (2017) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2017.2715400 · Cited by 42
“performance against two state-of-the-art methods: multi-view direct mapping using TV regularization (implemented following [1], [28]) and the single-view depth estimation using the network”
Rambach, J. R., Pagani, A., & Stricker, D. (2017) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar-adjunct.2017.42 · Cited by 34
“An alternative to that are Simultaneous Localization and Mapping (SLAM) approaches that extract features 3D and map their environment during tracking using structure from motion methods [7, 13].”
BackgroundEfficient Dense Point Cloud Object Reconstruction Using Deformation Vector Fields
Li, K., Pham, T. T., Zhan, H., & Reid, I. D. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-01258-8_31 · Cited by 39
“This branch of works try to find the geometry correspondence between views to recover geometry, such as SfM [13] and SLAM [25].”
BackgroundGSLAM: A General SLAM Framework and Benchmark
Zhao, Y., Xu, S., Bu, S., Jiang, H., & Han, P. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00120 · Cited by 41
“In recent years, SLAM technology develops rapidly and lots of monocular visual SLAM systems including key-point based [12, 37, 49], direct [15, 16, 53] and semi-direct methods [22, 23] are proposed.”
Visual-Based Semantic SLAM with Landmarks for Large-Scale Outdoor Environment
Zhao, Z., Mao, Y., Ding, Y., Ren, P., & Zheng, N. (2019) · China Symposium on Cognitive Computing and Hybrid Intelligence (CCHI) · doi:10.1109/cchi.2019.8901910 · Cited by 39
“DTAM [11], in which tracking is implemented by associating the input image with synthetic view images generated from the reconstructed map, and LSD-SLAM [12], which follows the idea from semidense VO [13], are the leading strategies in direct methods.”
MethodA scene-adaptive descriptor for visual SLAM-based locating applications in built environments
Xu, L., Feng, C., Kamat, V. R., & Menassa, C. C. (2020) · Automation in Construction · doi:10.1016/j.autcon.2019.103067 · Cited by 40
An Adaptive Visual Dynamic-SLAM Method Based on Fusing the Semantic Information
Jiao, J., Wang, C., Li, N., Deng, Z., & Xu, W. D. (2021) · IEEE Sensors Journal · doi:10.1109/jsen.2021.3051691 · Cited by 36
The Temporal Opportunist: Self-Supervised Multi-Frame Monocular Depth
Watson, J., Aodha, O. M., Prisacariu, V. A., Brostow, G., & Firman, M. (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/cvpr46437.2021.00122 · Cited by 473
PLJ-SLAM: Monocular Visual SLAM With Points, Lines, and Junctions of Coplanar Lines
Ren, G., Cao, Z., Liu, X., Tan, M. H., & Yu, J. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3185122 · Cited by 35
Santos, T. T., Koenigkan, L. V., Barbedo, J. G. A., & Rodrigues, G. C. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-16220-1_18 · Cited by 41
“These RGB-D devices are camera-projector triangulation-based systems assembled in a single device, also using coded-light to infer pixel correspondences [16].”
BackgroundInfluentialUAV‐based simultaneous localization and mapping in outdoor environments: A systematic scoping review
Wang, K., Kooistra, L., Pan, R., Wang, W., & Valente, J. R. P. (2024) · Journal of Field Robotics · doi:10.1002/rob.22325 · Cited by 45
“…such as EKF (PTAM; Klein & Murray, 2007), multistate constraint Kalman filter (MSCKF; Mourikis & Roumeliotis, 2007), dense tracking and mapping (Newcombe et al., 2011) tried to compare which method is better or combine the advantages of filtering methods and smoothing methods among them.”
Herrera-Granda, E. P., Torres, J. C., & Peluffo-Ordóñez, D. H. (2024) · Heliyon · doi:10.1016/j.heliyon.2024.e37356 · Cited by 49
Aggressive quadrotor flight using dense visual-inertial fusion
Ling, Y., Liu, T., & Shen, S. (2016) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2016.7487286 · Cited by 33
“With recent advances in high-performance mobile computing, direct dense methods have become popular [2, 3, 11]–[13].”
BackgroundExploiting Photometric Information for Planning Under Uncertainty
Costante, G., Delmerico, J., Werlberger, M., Valigi, P., & Scaramuzza, D. (2017) · Springer proceedings in advanced robotics · doi:10.1007/978-3-319-51532-8_7 · Cited by 38
GEN-SLAM: Generative Modeling for Monocular Simultaneous Localization and Mapping
Chakravarty, P., Narayanan, P., & Roussel, T. (2019) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2019.8793530 · Cited by 39
“Early work was based on single beam LIDAR and sonar, but vision based techniques [5], [6], [7] were popularized in the last decade.”
MethodDesaraju, V. R., Michael, N., Humenberger, M., Brockers, R., Weiß, S., Nash, J., & Matthies, L. H. (2015) · Autonomous Robots · doi:10.1007/s10514-015-9456-x · Cited by 42
“Newcombe et al (2011) introduce DTAM, a dense 3D reconstruction and mapping system that uses numerous images from the video stream to optimize the global model by minimizing a spatially regularized energy functional.”
MethodRobust Direct Visual Localisation using Normalised Information Distance
Pascoe, G., Maddern, W., & Newman, P. M. (2015) · British Machine Vision Conference (BMVC) · doi:10.5244/c.29.70 · Cited by 38
“Direct approaches have been successfully used for monocular SE(2) visual odometry at large scale [21] [26], SE(3) monocular localisation and mapping in indoor environments [28], and visual odometry with stereo [5] and RGB-D [38] sensors.”
BackgroundMethodInfluentialReal-time mesh-based scene estimation for aerial inspection
Teixeira, L., & Chli, M. (2016) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2016.7759714 · Cited by 39
“However, it was with DTAM [6] that the Computer Vision community has been”
InfluentialJust-in-Time Reconstruction: Inpainting Sparse Maps Using Single View Depth Predictors as Priors
Weerasekera, C. S., Dharmasiri, T., Garg, R., Drummond, T. W., & Reid, I. D. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8460549 · Cited by 37
“While dense SLAM methods like DTAM [1] are capable of real-time dense reconstruction of key-frames, methods using a point cloud representation for sparse or semi-dense mapping [2], [3], [4], [5] still remain a popular choice for large scale mapping and tracking.”
MethodStructure and Motion from Casual Videos
Zhang, Z., Cole, F. H., Li, Z., Rubinstein, M., Snavely, N., & Freeman, W. T. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19827-4_2 · Cited by 110
“Some visual SLAM methods instead seek to solve for dense or semi-dense depth per frame, and optimize directly based on image intensities rather than sparse correspondence [23, 14, 11, 10].”
BackgroundReal-Time Dense 3D Mapping of Underwater Environments
Wang, W., Joshi, B., Burgdorfer, N., Batsosc, K., Li, A. Q., Mordohaia, P., & Rekleitis, I. M. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160266 · Cited by 32
Shoukat, M. U., Yan, L., Deng, D., Imtiaz, M., Safdar, M., & Nawaz, S. A. (2024) · Advanced Engineering Informatics · doi:10.1016/j.aei.2024.102370 · Cited by 45
NIS-SLAM: Neural Implicit Semantic RGB-D SLAM for 3D Consistent Scene Understanding
Zhai, H., Huang, G., Hu, Q., Li, G., Bao, H., & Zhang, G. (2024) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2024.3456201 · Cited by 53
“DTAM [43] was the first dense SLAM system that leveraged pixel-level photometric information to perform camera pose tracking.”
Autonomous navigation of generic monocular quadcopter in natural environment
Bipin, K., Duggal, V., & Krishna, K. M. (2015) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2015.7139308 · Cited by 32
“At every controller update the initial state of the control problem is updated whereas update of target state depends upon arrival of new way-point from local planning module.”
BackgroundMethodFeature based graph-SLAM in structured environments
de la Puente, P., & Rodríguez-Losada, D. (2014) · Autonomous Robots · doi:10.1007/s10514-014-9386-z · Cited by 36
“Newcombe et al. (2011), for instance, pro-pose a regularized cost that promotes smooth reconstruction in dense visual SLAM.”
BackgroundFusion of Inertial and Visual Measurements for RGB-D SLAM on Mobile Devices
Brunetto, N., Salti, S., Fioraio, N., Cavallari, T., & Stefano, L. D. (2015) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2015.29 · Cited by 32
“On the other hand, monocular SLAM has reached a considerable maturity [7, 14] but still mandates massive parallelization by GPGPU to attain dense 3D reconstruction [18].”
BackgroundAn Integrated Visual Odometry System for Underwater Vehicles
Xu, Z., Haroutunian, M., Murphy, A. J., Neasham, J. A., & Norman, R. A. (2020) · IEEE Journal of Oceanic Engineering · doi:10.1109/joe.2020.3036710 · Cited by 68
“The well-known visual SLAM methods include Parallel81 Tracking And Mapping (PTAM) [14], Dense Tracking And Mapping (DTAM) [15] and large-82 scale direct (LSD-SLAM) [16].”
MethodOptical Flow-Based Stereo Visual Odometry With Dynamic Object Detection
Liu, Y., & Zhou, Z. (2022) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2022.3205015 · Cited by 34
“The other is the direct method [14], [15], [16] based on the photometric error.”
MethodNinjaDesc: Content-Concealing Visual Descriptors via Adversarial Learning
Ng, T., Kim, H. J., Lee, V. T., DeTone, D., Yang, T.-Y., Shen, T., Ilg, E., Balntas, V., Mikolajczyk, K., & Sweeney, C. (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/cvpr52688.2022.01246 · Cited by 42
“Local visual descriptors [7,13,56,73,75] are fundamental to a wide range of computer vision applications such as SLAM [15, 40, 42, 45], SfM [1, 65, 72], wide-baseline stereo [30,43], calibration [49], tracking [24,44,51], image retrieval [3, 4, 32, 46, 47, 67, 78, 79], and camera pose estimation [5,17,54,61,62,76,77].”
R3D3: Dense 3D Reconstruction of Dynamic Scenes from Multiple Cameras
Schmied, A., Fischer, T., Danelljan, M., Pollefeys, M., & Yu, F. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00298 · Cited by 63
“Early works utilize active depth sensors [36, 64] to alleviate the above-mentioned challenges.”
BackgroundLightNeuS: Neural Surface Reconstruction in Endoscopy Using Illumination Decline
Batlle, V. M., Montiel, J. M. M., Fua, P., & Tardós, J. D. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-43999-5_48 · Cited by 31
“Unfortunately, most of the previous 3D methods do not provide code [14,22], are not evaluated in biomedical settings [17,20], or do not report reconstruction accuracy [12,13].”
Simultaneous 3D tracking and reconstruction on a mobile phone
Prisacariu, V. A., Kähler, O., Murray, D. W., & Reid, I. D. (2013) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar.2013.6671768 · Cited by 32
“The established method for rendering a 3D shape represented in such a way is to use a raycasting algorithm [9].”
BackgroundMethodInfluentialAccurate keyframe selection and keypoint tracking for robust visual odometry
Fanfani, M., Bellavia, F., & Colombo, C. (2016) · Machine Vision and Applications · doi:10.1007/s00138-016-0793-3 · Cited by 34
“Camera tracking systems can also be characterized by the feature matching strategy adopted to detect and track keypoints across the image frames [6,11,17,18,25,29].”
BackgroundEnabling aggressive motion estimation at low-drift and accurate mapping in real-time
Zhang, J., & Singh, S. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989589 · Cited by 37
“For vision methods, in addition to stereo cameras [1], [2] and monocular cameras [3], [4] as common choices, RGB-D cameras have gained popularity in recent years.”
BackgroundImage Gradient-based Joint Direct Visual Odometry for Stereo Camera
Zhu, J. (2017) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2017/636 · Cited by 36
“[Newcombe et al., 2011] present a realtime dense tracking and mapping algorithm for a handheld monocular camera.”
BackgroundMethodResultInfluentialLearning Quadrangulated Patches for 3D Shape Parameterization and Completion
Sarkar, K., Varanasi, K., & Stricker, D. (2017) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2017.00051 · Cited by 41
3D mapping for multi hybrid robot cooperation
Surmann, H., Berninger, N., & Worst, R. (2017) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2017.8202217 · Cited by 36
“DTAM ([5]) also provides dense reconstructions in real-time.”
BackgroundMethodMaurer, D., Ju, Y. C., Breuß, M., & Bruhn, A. (2018) · International Journal of Computer Vision · doi:10.1007/s11263-018-1079-1 · Cited by 31
“This model is frequently used by recent stereo approaches, see e.g. Newcombe et al. (2011), Basha et al. (2012) and Semerjian (2014).”
MethodReview of vision-based Simultaneous Localization and Mapping
Li, A., Ruan, X., Huang, J., Zhu, X., & Wang, F. (2019) · IEEE Information Technology, Networking, Electronic and Automation Control Conference (ITNEC) · doi:10.1109/itnec.2019.8729285 · Cited by 34
“B. VSLAM system based on direct method DTAM [10] proposed in 2011, which is a monocular SLAM system based on dense direct method.”
MethodInfluentialDeepTAM: Deep Tracking and Mapping with Convolutional Neural Networks
Zhou, H., Ummenhofer, B., & Brox, T. (2019) · International Journal of Computer Vision · doi:10.1007/s11263-019-01221-0 · Cited by 41
“The most related work is DTAM (Newcombe et al, 2011).”
BackgroundMethodInfluentialPrivacy Preserving Visual SLAM
Shibuya, M., Sumikura, S., & Sakurada, K. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58542-6_7 · Cited by 39
Naveed, K., Anjum, M. L., Hussain, W., & Lee, D. (2022) · Autonomous Robots · doi:10.1007/s10514-022-10046-9 · Cited by 33
“Generally, multi-view geometry researchers produce a repetitive circular hand motion1 to gather parallax, which is good for reconstruction, while at the same time revisiting the area in order to reduce the uncertainty (Newcombe et al., 2011).”
BackgroundEnabling Aircraft Emergency Landings Using Active Visual Site Detection
Warren, M., Mejías, L., Yang, X., Arain, B., Gonzalez, F., & Upcroft, B. (2014) · Springer tracts in advanced robotics · doi:10.1007/978-3-319-07488-7_12 · Cited by 31
“As computational speed improves and algorithms become more efficient, near real-time implementations of SfM and Visual Odometry (VO) are now in the literature [7] [12] and can perform online on commodity hardware for small workspaces.”
Fast, robust, continuous monocular egomotion computation
Jaegle, A., Phillips, S., & Daniilidis, K. (2016) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2016.7487206 · Cited by 36
“Previous studies have used robust loss functions for monocular egomotion [21], visual SLAM [31], and RGBD odometry [32].”
MethodA Survey of Simultaneous Localization and Mapping with an Envision in 6G Wireless NetworksPreprint
Huang, B., Zhao, J., & Liu, J. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.05214 · Cited by 31
Dense Non-rigid Structure from Motion
Russell, C., Fayad, J. K., & Agapito, L. (2012) · International Conference on 3D Imaging, Modeling, Processing, Visualization & Transmission · doi:10.1109/3dimpvt.2012.70 · Cited by 30
“In this respect, they lie far behind their rigid Structure from Motion (SfM) counterparts which are even capable of a real time dense 3d reconstruction of static scenes that provides accurate depth information for every pixel in the image [19].”
BackgroundThree-Dimensional Traffic Scenes Simulation From Road Image Sequences
Li, Y., Liu, Y., Su, Y., Hua, G., & Zheng, N. (2016) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2015.2497408 · Cited by 30
“However, it is not only expensive and time consuming, but also unsafe to directly test on real traffic road.”
BackgroundEfficient Map Compression for Collaborative Visual SLAM
Opdenbosch, D. V., Aykut, T., Alt, N., & Steinbach, E. G. (2018) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2018.00114 · Cited by 30
“In general, (semi-) dense algorithms like DTAM [23] and LSD-SLAM [12] are more demanding in terms of computational complexity than their feature-based sparse counterparts.”
MethodSfMLearner++: Learning Monocular Depth & Ego-Motion Using Meaningful Geometric Constraints
Prasad, V., & Bhowmick, B. (2019) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2019.00226 · Cited by 33
Benchmark of Visual SLAM Algorithms: ORB-SLAM2 vs RTAB-Map
Ragot, N., Khemmar, R., Pokala, A., Rossi, R., & Ertaud, J.-Y. (2019) · International Conference on Emerging Security Technologies · doi:10.1109/est.2019.8806213 · Cited by 37
Self-supervised Visual-LiDAR Odometry with Flip Consistency
Li, B., Hu, M. H., Wang, S., Wang, L., & Gong, X. (2021) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv48630.2021.00389 · Cited by 33
“The direct methods [29, 14] minimize a photometric error that takes all pixels into account, assuming that scenes are stationary and of a Lambertian reflectance.”
MethodYang, A., Cui, C., Bârsan, I. A., Urtasun, R., & Wang, S. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9561481 · Cited by 37
Probabilistic Volumetric Fusion for Dense Monocular SLAM
Rosinol, A., Leonard, J. J., & Carlone, L. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00311 · Cited by 42
“For example, DTAM [12] achieves dense SLAM by using the same paradigm as the sparse PTAM [9], which tracked the camera pose first and then the depth, in a de-coupled fashion.”
MethodMIPS-Fusion: Multi-Implicit-Submaps for Scalable and Robust Online Neural RGB-D Reconstruction
Tang, Y., Zhang, J., Yu, Z., Wang, H., & Xu, K. (2023) · ACM Transactions on Graphics · doi:10.1145/3618363 · Cited by 47
Thermal-NeRF: Neural Radiance Fields from an Infrared Camera
Ye, T., Wu, Q., Deng, J., Liu, G., Liu, L., Xia, S., Pang, L., Yu, W., & Pei, L. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10802480 · Cited by 56
“Existing methods such as depth maps [14], point clouds [4], voxel grids [5], and meshes [6] have made significant contributions to this domain.”
A V-SLAM Guided and Portable System for Photogrammetric Applications
Torresani, A., Menna, F., Battisti, R., & Fabio Remondino, A. (2021) · Remote Sensing · doi:10.3390/rs13122351 · Cited by 29
“On the highest-level, V-SLAM algorithms are usually categorized in direct [39,40] and indirect [37,38,42–44,49] methods, and in filter-based [37] or graph-based methods [38,42–44,49].”
MethodElhashash, M., & Qin, R. (2022) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2022.03.018 · Cited by 32
Zhang, T., Zhang, L., Chen, Y., & Zhou, Y. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3213189 · Cited by 42
“In terms of the density of the constructed map, SLAM systems roughly fall into two categories, sparse ones [7]– [10] and dense ones [11]–[13].”
Prior depth-based multi-view stereo network for online 3D model reconstruction
Song, S., Giang, K. T., Kim, D., & Jo, S. (2022) · Pattern Recognition · doi:10.1016/j.patcog.2022.109198 · Cited by 29
“This system focused on accurate and detailed reconstruction instead of the fast depth computation considered in conventional motion stereo [7,8] .”
BackgroundMethodInfluentialLeonard, J. J., & Dellaert, F. (2014) · Cited by 29
3D Surface Reconstruction from Point-and-Line Cloud
Sugiura, T., Torii, A., & Okutomi, M. (2015) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2015.37 · Cited by 35
“Densification of 3D point cloud is possible when sequential images (video) are used [24] but prevents efficient reconstruction of 3D surface.”
BackgroundFeng, C., Chen, Z., Zhang, C., Hu, W., Li, B., & Lu, F. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3284479 · Cited by 29
“The cost volume C is constructed in the following manner: Given the previous frame I t − 1 , the current frame I t , and the estimated relative pose T , a set of ordered planes P is used as described in references [45] and [40].”
BackgroundFast and Accurate Large-Scale Stereo Reconstruction Using Variational Methods
Kuschk, G., & Cremers, D. (2013) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2013.96 · Cited by 34
“In [14], a quadratic relaxation between the convex regularizer and the non-convex data term was proposed for minimizing a Total Variation based optical flow energy functional and [8] used a similar approach for image driven and TV-based stereo estimation.”
MethodUnsupervised extrinsic calibration of depth sensors in dynamic scenes
Miller, S., Teichman, A., & Thrun, S. (2013) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2013.6696737 · Cited by 28
“And while state of the art reconstruction techniques such as DTAM [15] require a continuous trajectory rather than disparate viewpoints, the task of finding loop closures is quite similar to our own.”
BackgroundPoint-Based 3D Reconstruction of Thin Objects
Ummenhofer, B., & Brox, T. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.124 · Cited by 28
“[19] to improve the initial camera parameters.”
Manhattan and Piecewise-Planar Constraints for Dense Monocular Mapping
Concha, A., Hussain, W., Montano, L., & Civera, J. (2014) · Robotics: Science and Systems · doi:10.15607/rss.2014.x.016 · Cited by 28
“DTAM can also be inaccurate for low-parallax camera motions and non-Lambertian effects –this latest limitation is mentioned in the original paper [17].”
BackgroundMethodInfluentialMarker-based non-overlapping camera calibration methods with additional support camera views
Zhao, F., Tamaki, T., Kurita, T., Raytchev, B., & Kaneda, K. (2018) · Image and Vision Computing · doi:10.1016/j.imavis.2017.12.006 · Cited by 29
Chen, Y., Zhou, Y., Lv, Q., & Deveerasetty, K. K. (2018) · IEEE International Conference on Information and Automation (ICIA) · doi:10.1109/icinfa.2018.8812387 · Cited by 38
“And in DTAM, the depth of each pixel is executed inverse-depth calculation and continuous optimization on the premise of the photometric invariance.”
BackgroundInfluential∇SLAM: Dense SLAM meets Automatic Differentiation
Murthy, J. K., Iyer, G., & Paull, L. (2020) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra40945.2020.9197519 · Cited by 38
“In dense SLAM systems [7], [8], each pixel in each image contributes to the dense 3D map.”
BackgroundHardware implementation of SLAM algorithms: a survey on implementation approaches and platforms
Eyvazpour, R., Shoaran, M., & Karimian, G. (2022) · Artificial Intelligence Review · doi:10.1007/s10462-022-10310-5 · Cited by 37
Robustness to lighting variations: An RGB-D indoor visual odometry using line segments
Lu, Y., & Song, D. (2015) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2015.7353447 · Cited by 28
“In the meantime, dense approaches [7] are demonstrated to permit superior accuracy than sparse feature-based algorithms, though they rely on the constant brightness assumption and need GPU acceleration for realtime purpose.”
BackgroundSurface Reconstruction via Fusing Sparse-Sequence of Depth Images
Yang, L., Yan, Q., Fu, Y., & Xiao, C. (2017) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2017.2657766 · Cited by 30
“Frame-to-model ICP algorithm is sufficient for aligning an indoor scale target [7], [21].”
BackgroundDF-VO: What Should Be Learnt for Visual Odometry?
Zhan, H., Weerasekera, C. S., Bian, J.-W., Garg, R., & Reid, I. (2021) · arXiv · doi:10.48550/arxiv.2103.00933 · Cited by 31
“Two dominant methods for geometry-based VO/SLAM are feature-based (Geiger et al. 2011; Klein & Murray 2007; Mur-Artal & Tard´os 2016) and direct methods (Engel et al. 2017; Newcombe et al. 2011).”
Monocular Visual-Inertial Odometry with Planar Regularities
Chen, C., Geneva, P., Peng, Y., Lee, W., & Huang, G. P. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160620 · Cited by 40
HI-SLAM: Monocular Real-Time Dense Mapping With Hybrid Implicit Fields
Zhang, W., Sun, T., Wang, S., Cheng, Q., & Norbert Haala, A. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3347131 · Cited by 34
“DTAM [15] pioneered one of the first real-time dense SLAM systems by parallelizing depth computing on GPU.”
BackgroundWang, C., Yuan, J., & Xie, L. (2017) · International Conference on Advanced Robotics (ICAR) · doi:10.1109/icar.2017.8023500 · Cited by 32
“Being nonrobustness to illumination changes is one of the potential problems of matching pixel intensities [14].”
BackgroundMonocular Vision-Based Obstacle Detection and Avoidance for a Multicopter
Chen, H.-C. (2019) · IEEE Access · doi:10.1109/access.2019.2953954 · Cited by 29
“There are other studies that obtain depth based on geometric methods [12], [13], [14].”
BackgroundChen, C., Zhu, H. J., Wang, L., & Liu, Y. (2019) · IEEE Access · doi:10.1109/access.2019.2961266 · Cited by 31
“[18] R. A. Newcombe, S. J. Lovegrove, and A. J. Davison, ‘‘DTAM: Dense tracking and mapping in real-time,’’ in Proc.”
BackgroundDirect Visual-Inertial Ego-Motion Estimation Via Iterated Extended Kalman Filter
Zhong, S., & Chirarattananon, P. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.2968071 · Cited by 28
“Nevertheless, for map-based methods, intensive computation is required for the generation of dense depth-map [14].”
MethodXu, C., Liu, Z., & Li, Z. (2021) · Remote Sensing · doi:10.3390/rs13040772 · Cited by 29
Paton, M., & Košecká, J. (2012) · Canadian Conference on Computer and Robot Vision · doi:10.1109/crv.2012.11 · Cited by 27
“In addition to visual odometry and sparse feature models, works of [15], [16] focus on dense 3D reconstruction.”
BackgroundModel Learning and Real-Time Tracking Using Multi-Resolution Surfel Maps
Stückler, J., & Behnke, S. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v26i1.8388 · Cited by 27
Model-based motion blur estimation for the improvement of motion tracking
Seibold, C., Hilsmann, A., & Eisert, P. (2017) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2017.03.005 · Cited by 27
Photometric patch-based visual-inertial odometry
Zheng, X., Moratto, Z., Li, M., & Mourikis, A. I. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989372 · Cited by 31
“Depending on the type of image regions used in the algorithm, these can be further divided into dense methods, where the entire image is used [11], semi-dense methods, where only regions with large gradient magnitude are used [12], [13], and patch-based methods, where regions around extracted point-features are used […”
MethodSelection and Compression of Local Binary Features for Remote Visual SLAM
Opdenbosch, D. V., Oelsch, M., Garcea, A., Aykut, T., & Steinbach, E. G. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8463202 · Cited by 27
“On the other hand, dense (DTAM [6]) or semi-dense approaches (LSD-SLAM [7]) try to minimize the photometric error between frames using the raw pixel intensities.”
BackgroundRobotics: Science and Systems XI
(2015) · doi:10.15607/rss.2015.xi · Cited by 28
Efficient edge-aware surface mesh reconstruction for urban scenes
Bódis-Szomorú, A., Riemenschneider, H., & Gool, L. V. (2016) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2016.06.002 · Cited by 29
3-D Tracking for Augmented Reality Using Combined Region and Dense Cues in Endoscopic Surgery
Wang, R., Mei, Z., Meng, X., Geng, Z., & Wang, F. (2017) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2017.2770214 · Cited by 27
“There are mainly two types of SLAM techniques: parallel tracking and mapping (PTAM) [13] using sparse feature points and dense tracking and mapping (DTAM) [14] using all the pixels in the image.”
MethodLi, S., Ren, B., Liu, Y., Cheng, M., Frost, D., & Prisacariu, V. A. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8461003 · Cited by 27
“2) Direct Point-Based methods [3] [10] [19] [20] [11] [21]: Compared with the former one, Direct Point-Based methods skip the procedure of feature extraction and feature matching.”
MethodDifferential dynamic programming for optimal estimation
Kobilarov, M., Ta, D.-N., & Dellaert, F. (2015) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2015.7139279 · Cited by 25
“Recent estimation methods based on related ideas have focused on fast performance for real-time navigation [10], [25], [24], [23], [31], large-scale problems with massive amount of sensor data [1], [6], [33], [34], [22], [11], and long-term navigation [35], [36], [8], [9], [32].”
BackgroundOcclusion Matting: Realistic Occlusion Handling for Augmented Reality Applications
Hebborn, A. K., Höhner, N., & Müller, S. (2017) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar.2017.23 · Cited by 26
“Today’s reconstruction algorithms [20, 35] are able to reconstruct a semi dense and accurate model, while the boundaries, however, are often still noisy.”
BackgroundVisual Semantic Landmark-Based Robust Mapping and Localization for Autonomous Indoor Parking
Zhao, J., Huang, Y., He, X., Zhang, S., Ye, C., Feng, T., & Xiong, L. (2019) · Sensors · doi:10.3390/s19010161 · Cited by 29
“Moreover, a high frame rate is also required since brightness consistency is crucial for estimating the depth accurately [18].”
BackgroundA Framework for the Volumetric Integration of Depth ImagesPreprint
Prisacariu, V. A., Kähler, O., Cheng, M., Ren, C. Y., Valentin, J. P. C., Torr, P. H. S., Reid, I. D., & Murray, D. W. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1410.0925 · Cited by 42
“However, monocular SLAM has since also been tackled with very similar approaches [8, 11].”
BackgroundMethod
Metadata
- Authors
- Newcombe, R. A., Lovegrove, S. J., & Davison, A. J.
- Affiliations
- Imperial College London (GB)
- ORCID
- Davison, Andrew J.
- Venue
- 2011 International Conference on Computer Vision
- Publisher
- IEEE
- Published
- 2011-11
- Pages
- 2320-2327
- OpenAlex
- W2108134361
- Cited by
- 2,078
- References
- 17
- Citations by year
- 2026: 49 · 2025: 144 · 2024: 163 · 2023: 140 · 2022: 141 · 2021: 166 · 2020: 179 · 2019: 199 · 2018: 183 · 2017: 176 · 2016: 136 · 2015: 151 · 2014: 102 · 2013: 92 · 2012: 51
- FWCI
- 154.56
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE/CVF International Conference on Computer Vision (ICCV) (publishedVersion)
- 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 › Earth and Planetary Sciences › Geology › 3D Surveying and Cultural Heritage
- Keywords
- dense 3D reconstruction, real-time mapping, depth images, non-convex optimization, energy minimization, image registration, GPU parallelization, augmented reality, handheld camera
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.