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 701–800 of 1,004
IVPR: An Instant Visual Place Recognition Approach Based on Structural Lines in Manhattan World
Pei, L., Liu, K., Zou, D., Li, T., Wu, Q., Zhu, Y., Li, Y., He, Z., Chen, Y., & Sartori, D. (2019) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2019.2941359 · Cited by 23
“In general, there are two perspectives to classify SLAM approaches: sparse [1]/semidense [2]/dense [3] solutions and direct [4]/semidirect [5]/indirect [6] methods.”
BackgroundDiscontinuous and Smooth Depth Completion With Binary Anisotropic Diffusion Tensor
Yao, Y., Roxas, M., Ishikawa, R., Ando, S., Shimamura, J., & Oishi, T. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.3005890 · Cited by 22
“Moreover, we use inverse depth d̄−1 instead of depth d̄ to balance the contribution of near and far depth as was done in [19].”
MethodMOLTR: Multiple Object Localization, Tracking and Reconstruction From Monocular RGB Videos
Li, K., Rezatofighi, H., & Reid, I. D. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3061080 · Cited by 21
“The increasing computation power enables dense SLAM systems [3]–[5], where the reconstruction is composed of dense surfels or Truncated Signed Distance Function (TSDF).”
BackgroundKühne, J., Magno, M., & Benini, L. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3406948 · Cited by 33
“The detection of features on image frames is often performed with a layered approach, where first a fast algorithm is run to find candidate features.”
Autonomous control of unmanned aerial vehicles: applications, requirements, challenges
Küçükerdem, H., Yılmaz, C., Kahraman, H. T., & Sönmez, Y. (2025) · Cluster Computing · doi:10.1007/s10586-025-05418-6 · Cited by 30
“…a nonlinear image alignment problem [165] DTAM, a real-time SLAM system employing linear techniques for camera motion estimation, has been utilized [166] A novel approach has been introduced to the SVO system for aerial vehicle attitude estimation [167] Proportional-Integral-Derivative (PID),…”
Large Scale Dense Visual Inertial SLAM
Ma, L., Falquez, J. M., McGuire, S., & Sibley, G. (2016) · Springer tracts in advanced robotics · doi:10.1007/978-3-319-27702-8_10 · Cited by 19
Self-Supervised Learning of Depth and Motion Under Photometric Inconsistency
Shen, T., Zhou, L., Luo, Z., Yao, Y., Li, S., Zhang, J., Fang, T., & Quan, L. (2019) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2019.00499 · Cited by 17
“The self-supervised depth and motion learning framework derives from direct SLAM methods [9, 10, 35].”
MethodA Robust Stereo Feature-aided Semi-direct SLAM System
Zhao, X., Liu, L., Zheng, R., Ye, W., & Liu, Y. (2020) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2020.103597 · Cited by 19
A SLAM System Based on RGBD Image and Point-Line Feature
Li, D., Liu, S., Xiang, W., Tan, Q., Yuan, K., Zhang, Z., & Hu, Y. (2021) · IEEE Access · doi:10.1109/access.2021.3049467 · Cited by 21
“Newcombe’s DTAM (Dense tracking and mapping) [23], one of the representatives of the direct method, can achieve real-time recovery of 3D models, and has high robustness in image blur and feature loss.”
Sensor Deployment for Visual 3D Perception: A Perspective of Information Gains
An, Q., Wang, Y., & Shen, Y. (2021) · IEEE Sensors Journal · doi:10.1109/jsen.2021.3050325 · Cited by 19
“Different from indirect methods that extract image features, direct methods utilize pixels to build a dense or semi-dense map and adopt specific data structures like the pose graph [21] or computing acceleration tools like GPU [22] to realize the real-time performance.”
MethodUncLe-SLAM: Uncertainty Learning for Dense Neural SLAM
Sandström, E., Ta, K., Gool, L. V., & Oswald, M. R. (2023) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw60793.2023.00488 · Cited by 20
“Further advancements led to dense SLAM, such as [44, 61, 67, 81], which can also handle loop closures such as BundleFusion [15].”
BackgroundA New Era of Indoor Scene Reconstruction: A Survey
Wang, H., & Li, M. (2024) · IEEE Access · doi:10.1109/access.2024.3440260 · Cited by 21
A high-performance system-on-chip architecture for direct tracking for SLAM
Boikos, K., & Bouganis, C.-S. (2017) · International Conference on Field Programmable Logic and Applications (FPL) · doi:10.23919/fpl.2017.8056831 · Cited by 23
“On the other hand, they are highly complex and data-intensive and require at least high-end desktop-grade hardware and often GPU acceleration [1] to process all of the available information at a framerate of 30 frames/s or more.”
BackgroundReal-Time Cloud Visual Simultaneous Localization and Mapping for Indoor Service Robots
Zheng, Y., Chen, S., & Cheng, H. (2020) · IEEE Access · doi:10.1109/access.2020.2966757 · Cited by 20
Vivid: Augmenting Vision-Based Indoor Navigation System With Edge Computing
Zhao, W., Xu, L., Qi, B., Hu, J. S., Wang, T., & Runge, T. M. (2020) · IEEE Access · doi:10.1109/access.2020.2978123 · Cited by 26
“DTAM [32] needs GPU to run in real time, LSD-SLAM [9] and ORB-SLAM [6] work on computer CPU.”
MethodImproving Dense Mapping for Mobile Robots in Dynamic Environments Based on Semantic Information
Cheng, J., Wang, C., Mai, X., Min, Z., & Meng, M. Q. ‐. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.3023696 · Cited by 19
“Nowadays, with the development of visual simultaneous localization and mapping (visual SLAM), the mapping process [1] has become more and more promising.”
BackgroundDeepSurfels: Learning Online Appearance Fusion
Mihajlović, M., Weder, S., Pollefeys, M., & Oswald, M. R. (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/cvpr46437.2021.01429 · Cited by 19
“Appearance mapping is often treated as a separate post-processing step that follows 3D surface reconstruction and is usually approached using batch-based optimization methods [19, 20, 90, 25] that are unsuitable for many applications that do not have access to the entire dataset at processing time, for instance, robot…”
BackgroundToward 3D object reconstruction from stereo images
Xie, H., Yao, H., Zhou, S., Zhang, S., Tong, X., & Sun, W. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.07.089 · Cited by 19
Layout aware visual tracking and mapping
Salas, M., Hussain, W., Concha, A., Montano, L., Civera, J., & Montiel, J. M. M. (2015) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2015.7353367 · Cited by 19
“We utilize the 3D box layout to select images coming from the same room, which improves the DTAM reconstruction in multi-room scenario.”
BackgroundMethodInfluentialZheng, J., Chen, T., He, J., Wang, Z., & Gao, B. (2024) · Energies · doi:10.3390/en17164106 · Cited by 18
MG-SLAM: Structure Gaussian Splatting SLAM With Manhattan World Hypothesis
Liu, S., Deng, T., Zhou, H., Li, L., Wang, H., Wang, D., & Li, M. (2025) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2025.3575772 · Cited by 45
“Learning-based dense SLAM methods, particularly neural radiance field (NeRF) approaches [1]–[8], have demonstrated remarkable improvements in capturing dense photometric information and providing accurate global reconstruction over traditional systems based on sparse point clouds [9]–[15].”
Simultaneous Super-Resolution of Depth and Images Using a Single Camera
Lee, H. S., & Lee, K. M. (2013) · 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.2013.43 · Cited by 18
“After the initial depth map is created, the image registration method similar to the 2.5D image registration in [10] is used between the input frame and the pre-warped image from the estimated high-resolutionimage and depth map to estimate a new camera pose P J +1 as: The optimization of this…”
BackgroundMethodSemantic Texture for Robust Dense Tracking
Czarnowski, J., Leutenegger, S., & Davison, A. J. (2017) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2017.105 · Cited by 18
“Dense visual SLAM [15] involves incrementally reconstructing the whole appearance of a scene rather reducing it to sparse features, and this approach is fundamental if we want scene models with general, generative representations.”
BackgroundMethodSelf-citationVolumetric Bundle Adjustment for Online Photorealistic Scene Capture
Clark, R. (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/cvpr52688.2022.00603 · Cited by 18
“This is typically done by constructing a photometric cost-volume and extracting per-pixel depth values that minimize this cost [14].”
BackgroundA Review of Common Techniques for Visual Simultaneous Localization and Mapping
Dai, Y., Wu, J., & Wang, D. (2023) · Journal of Robotics · doi:10.1155/2023/8872822 · Cited by 25
Wang, Z., Qin, C., & Mu, X. (2024) · Sensors · doi:10.3390/s24061937 · Cited by 24
“These methods, like DTAM [28], SVO [29], and LSD-SLAM [30], directly employ image brightness information to estimate the current location of the body.”
MethodApplication of Event Cameras and Neuromorphic Computing to VSLAM: A Survey
Tenzin, S., Rassau, A., & Chai, D. (2024) · Biomimetics · doi:10.3390/biomimetics9070444 · Cited by 32
“However, most of the successfully developed algorithms such as MonoSLAM [42], PTAM [43], DTAM [44], and SLAM++ [45] have been developed for stationary environments, where it is assumed that the camera is the sole moving item in a static environment.”
MethodDSM Accuracy Evaluation for the ISPRS Commission I Image Matching Benchmark
Kuschk, G., d’Angelo, P., Qin, R., Poli, D., Reinartz, P. H., & Cremers, D. (2014) · The international archives of the photogrammetry, remote sensing and spatial information sciences/International archives of the photogrammetry, remote sensing and spatial information sciences · doi:10.5194/isprsarchives-xl-1-195-2014 · Cited by 18
“…et al., 2009), a quadratic relaxation between the convex regularizer∇u(x) and the non-convex data termC(x, u(x)) was proposed for minimizing a Total Variation based optical flow energy functional and (Newcombe et al., 2011) used a similar approach for image driven and TV-based stereo estimation.”
MethodInfluentialAppearance-based indoor localization: A comparison of patch descriptor performance
Rivera-Rubio, J., Alexiou, I., & Bharath, A. A. (2015) · Pattern Recognition Letters · doi:10.1016/j.patrec.2015.03.003 · Cited by 23
“Finally, we applied one implementation of the SLAM to this ataset, at the same frame resolution as for the appearance-based Method Metric Error summary (cm) AUC (”
BackgroundInfluential3D Time-Lapse Reconstruction from Internet Photos
Martin-Brualla, R., Gallup, D., & Seitz, S. M. (2015) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.157 · Cited by 18
Mobile phone and cloud — A dream team for 3D reconstruction
Locher, A., Perdoch, M., Riemenschneider, H., & Gool, L. V. (2016) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2016.7477629 · Cited by 19
“In contrast to these feature-based tracking methods, DTAM [25] uses whole image alignment, allowing a very robust tracking.”
MethodFeature-based RGB-D camera pose optimization for real-time 3D reconstruction
Wang, C., & Guo, X. (2017) · Computational Visual Media · doi:10.1007/s41095-016-0072-2 · Cited by 18
Monocular visual odometry: Sparse joint optimisation or dense alternation?
Platinsky, L., Davison, A. J., & Leutenegger, S. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989599 · Cited by 18
“However, recently we have seen a strong shift towards semi-dense and fully dense VO algorithms [10], [11] which perform real-time estimation of the locations of many thousands of points and use photometric alignment against all of them to track camera motion.”
BackgroundMethodSelf-citationApplication of Augmented Reality and Robotic Technology in Broadcasting: A Survey
Yan, D., & Hu, H. (2017) · Robotics · doi:10.3390/robotics6030018 · Cited by 19
“DTAM belongs to key-frame-based SLAM, and allows every captured feature contributing tracking performance [67].”
BackgroundMulti-modal and inertial sensor solutions for navigation-type factor graphs
Fourie, D. (2017) · Woods Hole Open Access Server (WHOAS) · doi:10.1575/1912/9305 · Cited by 18
“[173] and [171] demonstrated dense tracking and mapping (DTAM), which was able to resolve both location and mapping using only a RGB-D sensor.”
BackgroundHuang, K., Wang, Y., & Kneip, L. (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.2019.01299 · Cited by 18
“While direct sparse [2], semi-dense [3], and dense methods [26] have already been presented, the more traditional way of sparse feature correspondence-based localization and mapping continues to be of high importance.”
BackgroundDenseFusion: Large-Scale Online Dense Pointcloud and DSM Mapping for UAVs
Chen, L., Zhao, Y., Xu, S., Bu, S., Han, P., & Wan, G. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341413 · Cited by 25
“As for direct monocular SLAM, DTAM [14] achieves real-time dense reconstruction on GPU by using short baseline stereo matching and a regularization method, which make the depth estimation of low texture area smoother.”
MethodGao, H., Liu, X., Qu, M., & Huang, S. (2021) · Applied Sciences · doi:10.3390/app11125383 · Cited by 20
“It has been widely applied in AR (augmented reality) [1], robotics navigation [2], and self-driving cars [3] for generating high-quality depth from color instead of using expensive LiDAR sensors.”
MethodRobust self-supervised monocular visual odometry based on prediction-update pose estimation network
Xiu, H., Liang, Y., Zeng, H., Li, Q., Liu, H., Fan, B., & Li, C. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.105481 · Cited by 19
Autonomous Navigation in Complex Indoor and Outdoor Environments with Micro Aerial Vehicles
Shen, S. (2014) · ScholarlyCommons (University of Pennsylvania) · Cited by 26
Monocular thermal SLAM with neural radiance fields for 3D scene reconstruction
Wu, Y., Wang, L., Zhang, L., Chen, M., Zhao, W., Zheng, D., & Cai, Y. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.129041 · Cited by 17
Realtime Edge Based Visual Inertial Odometry for MAV Teleoperation in Indoor Environments
Tarrio, J., & Pedre, S. (2017) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-017-0670-y · Cited by 18
“Even though some of these algorithms are able to run in real time in modern cpu+gpu equipment [25, 27] their computational cost is still too high for embedded solutions.”
BackgroundLoRa Backscatter Assisted State Estimator for Micro Aerial Vehicles with Online Initialization
Zhang, S., Wang, W., Zhang, N., & Jiang, T. (2021) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2021.3063850 · Cited by 22
“Although vision-based direct methods [50], [51] that work with all the raw pixel information in images have better performance in dealing with textureless scenes, they require high computing power (GPUs) to achieve real-time processing, which is unavailable for payload-limited MAVs.”
Dynamic visual simultaneous localization and mapping based on semantic segmentation module
Jin, J., Jiang, X., Yu, C., Zhao, L., & Tang, Z. (2023) · Applied Intelligence · doi:10.1007/s10489-023-04531-6 · Cited by 19
“In contrast to featurebased methods, direct methods do not rely on feature points, such as LSD-SLAM [5], SVO [15], DTAM [16], and DSO [6].”
MethodAutomatic Multi-Camera Calibration and Refinement Method in Road Scene for Self-Driving Car
Li, J., Pi, J., Wei, P., Luo, Z., & Yan, G. (2023) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2023.3323665 · Cited by 19
“Nowadays more and more researchers are willing to adopt direct methods rather than feature-point based ones to recover camera poses from images, especially in the field of SLAM [13], [14], [15], [16], [17].”
BackgroundUser friendly SLAM initialization
Mulloni, A., Ramachandran, M., Reitmayr, G., Wägner, D., Grasset, R., & Diaz, S. (2013) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar.2013.6671775 · Cited by 22
“Our system uses similar formulation to the depth-bundles described by Pietzsch [16], where only the inverse depth of each feature is estimated.”
MethodReal-time height map fusion using differentiable rendering
Zienkiewicz, J., Davison, A. J., & Leutenegger, S. (2016) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2016.7759630 · Cited by 21
“Monocular RGB methods like DTAM [19] have generally not been nearly as robust.”
MethodSelf-citationEdge alignment-based visual–inertial fusion for tracking of aggressive motions
Ling, Y., Kuse, M., & Shen, S. (2017) · Autonomous Robots · doi:10.1007/s10514-017-9642-0 · Cited by 16
“Newcombe et al. (2011) present a system for real-time camera tracking and reconstruction using current commodity GPU hardware.”
BackgroundMethodInfluentialLight condition invariant visual SLAM via entropy based image fusion
Kim, J.-W., & Kim, A. (2017) · International Conference on Ubiquitous Robots and Ambient Intelligence · doi:10.1109/urai.2017.7992661 · Cited by 17
“DTAM [3] provides accurate and dense 3-D maps, and LSD-SLAM [4] provides semi-dense 3-D maps in large scale environments.”
BackgroundRobust Stereo Visual Odometry Using Improved RANSAC-Based Methods for Mobile Robot Localization
Liu, Y., Gu, Y., Li, J., & Zhang, X. (2017) · Sensors · doi:10.3390/s17102339 · Cited by 16
A voxelize structured refinement method for registration of point clouds from Kinect sensors
Özbay, E., & Çınar, A. (2018) · Engineering Science and Technology an International Journal · doi:10.1016/j.jestch.2018.09.012 · Cited by 20
Combining Depth Fusion and Photometric Stereo for Fine-Detailed 3D Models
Bylow, E., Maier, R. L., Kahl, F., & Olsson, C. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-20205-7_22 · Cited by 17
“Classical algorithms include Structure from Motion [1, 2] which yields sparse point clouds and multiple-view stereo [3, 4] which generates dense reconstructions.”
BackgroundContinuous Direct Sparse Visual Odometry from RGB-D Images
Jadidi, M. G., Clark, W., Bloch, A. M., Eustice, R. M., & Grizzle, J. W. (2019) · Robotics: Science and Systems · doi:10.15607/rss.2019.xv.044 · Cited by 26
EMPNet: Neural Localisation and Mapping Using Embedded Memory Points
Avraham, G., Zuo, Y., Dharmasiri, T., & Drummond, T. W. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00821 · Cited by 17
“DeepTAM [42] reformulated the Dense Tracking and Mapping (DTAM) approach of [31] as a learning problem.”
BackgroundMethodA Hybrid Sparse-Dense Monocular SLAM System for Autonomous Driving
Gallagher, L., Kumar, V. R., Yogamani, S. K., & McDonald, J. (2021) · European Conference on Mobile Robots · doi:10.1109/ecmr50962.2021.9568797 · Cited by 23
Chen, T., Pu, F., Chen, H., & Liu, Z. (2022) · Remote Sensing · doi:10.3390/rs14092033 · Cited by 18
Aasim, W. F. A. W., Okasha, M., & Faris, W. F. (2022) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-022-01643-y · Cited by 18
“Some well-known earlier SLAM solutions utilising Direct methods are LSD-SLAM [21] and DTAM [23].”
MethodCao, Y., Zhang, X.-S., Luo, F., Peng, P., Lin, C., Yang, K.-F., & Li, Y. (2022) · Pattern Recognition · doi:10.1016/j.patcog.2022.109262 · Cited by 20
“, robotics navigation [1] and augmented reality [2].”
BackgroundMethodReal-time localization and dense mapping in underwater environments from a monocular sequence
Concha, A., Drews, P., Campos, M. F. M., & Civera, J. (2015) · MTS/IEEE OCEANS Conference (OCEANS) · doi:10.1109/oceans-genova.2015.7271476 · Cited by 17
“It is worth remarking that, as reported in [19], [20], dense mapping can be greatly speeded up with GPU programming.”
BackgroundMethodZhang, Y., Luo, C., & Liu, J. (2012) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/2370216.2370285 · Cited by 17
“Comparing to dense tracking and mapping with 2D cameras [20, 14, 2], RGB-D cameras can be more robust to lighting and image features due to the additional depth images.”
BackgroundIncremental light bundle adjustment for structure from motion and robotics
Indelman, V., Roberts, R. B., & Dellaert, F. (2015) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2015.03.009 · Cited by 17
“[38] presented an approach for real time camera tracking and dense reconstruction.”
MethodStereo and kinect fusion for continuous 3D reconstruction and visual odometry
Yılmaz, Ö., & Karakus, F. (2013) · International Conference on Electronics, Computer and Computation · doi:10.1109/icecco.2013.6718242 · Cited by 17
“Many robust solutions have been proposed in the recent years for 3D reconstruction of a scene using active (Kinect, ToF cameras) or passive (stereo) visual sensors [1-7].”
BackgroundRoom layout estimation from rapid omnidirectional exploration
Lukierski, R., Leutenegger, S., & Davison, A. J. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989747 · Cited by 17
“This is performed using a cost volume/plane sweep approach, as described in [11], [6].”
MethodSelf-citationImPACT-TRC Thin Serpentine Robot Platform for Urban Search and Rescue
Konyo, M., Ambe, Y., Nagano, H., Yamauchi, Y., Tadokoro, S., Bando, Y., Itoyama, K., Okuno, H. G., Okatani, T., Shimizu, K., et al. (2019) · Springer tracts in advanced robotics · doi:10.1007/978-3-030-05321-5_2 · Cited by 16
Neri, A., Penza, V., Baldini, C., & Mattos, L. S. (2025) · Computerized Medical Imaging and Graphics · doi:10.1016/j.compmedimag.2025.102616 · Cited by 20
Scalable Structure from Motion for Densely Sampled Videos
Resch, B., & Lensch, H. P. A. (2016) · Cited by 17
6D Dynamic Camera Relocalization from Single Reference Image
Feng, W., Tian, F.-P., Zhang, Q., & Sun, J. (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.2016.439 · Cited by 16
“Camera pose registration and relocalization is an essential problem in computer vision and robotics, fundamentally supporting a number of important real-world applications, such as structure-from-motion (SfM) [22, 37, 16], 3D tracking and mapping [24, 17], monocular SLAM [2, 34], and scene change detection [7, 5].”
BackgroundMethodAn Accurate, Robust Visual Odometry and Detail-Preserving Reconstruction System
Gong, X., Liu, Y., Wu, Q., Huang, J., Zong, H., & Wang, J. (2020) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2020.3017886 · Cited by 18
“DTAM [14] is a dense tracking and mapping system, which directly estimate the pose of the camera without using sparse features.”
Hierarchical Temporal Context Learning for Camera-Based Semantic Scene Completion
Li, B., Deng, J., Zhang, W., Liang, Z., Du, D., Jin, X., & Zeng, W. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73235-5_8 · Cited by 25
Scene Representations for Robotic Spatial Perception
Mascaro, R., & Chli, M. (2024) · Annual Review of Control Robotics and Autonomous Systems · doi:10.1146/annurev-control-040423-030709 · Cited by 34
Reliable kinect-based navigation in large indoor environments
Jalobeanu, M., Shirakyan, G., Parent, G., Kikkeri, H., Peasley, B., & Feniello, A. (2015) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2015.7139225 · Cited by 17
“An RGB-only variant called DTAM that also uses frame-to-model matching but is based on dense RGB data and a photometric distance was also proposed by Newcombe et al. [6].”
MethodMonocular scene flow estimation via variational method
Xiao, D., Yang, Q., Yang, B., & Wei, W. (2015) · Multimedia Tools and Applications · doi:10.1007/s11042-015-3091-6 · Cited by 16
“The conception of inverse depth came from Civer’s study about mono SLAM [7] and Richard emphasized it in his DTAM system [21].”
BackgroundVisualization Techniques for Augmented Reality in Endoscopic Surgery
Wang, R., Geng, Z., Zhang, Z., & Pei, R. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-43775-0_12 · Cited by 16
“In the reconstruction part, we use 5 captured endoscopic images and adopt the algorithm in [16] to get a reconstructed depth map and point cloud.”
Method$$\pi $$Match: Monocular vSLAM and Piecewise Planar Reconstruction Using Fast Plane Correspondences
Raposo, C., & Barreto, J. P. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46484-8_23 · Cited by 16
“While feature-based methods work on top of extracted features and are usually robust to outliers by applying RANSAC-based schemes [11], direct methods perform whole image alignment and cannot handle outliers [19, 4, 2].”
MethodLarge-Scale Tracking for Images With Few Textures
Lu, G., Nie, L., Sorensen, S., & Kambhamettu, C. (2017) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2017.2731044 · Cited by 16
“Depth information with RGB pixels is also applied for dense 3D tracking without feature extraction [55].”
Onboard real-time dense reconstruction of large-scale environments for UAV
Vempati, A. S., Gilitschenski, I., Nieto, J. I., Beardsley, P. A., & Siegwart, R. (2017) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2017.8206189 · Cited by 18
“DTAM [2] is one of the first SLAM systems that uses dense methods for scene reconstruction from an RGB camera in real-time using a parallelized framework based on a GPU.”
ENG: End-to-End Neural Geometry for Robust Depth and Pose Estimation Using CNNs
Dharmasiri, T., Spek, A., & Drummond, T. W. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-20887-5_39 · Cited by 20
A Sparse Visual Odometry Technique Based on Pose Adjustment With Keyframe Matching
Lin, H., & Hsu, J.-L. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.3015922 · Cited by 17
“The early successful works, including Kinect-Fusion [28] and ElasticFusion [29], use the depth information and truncated signed distance function (TSDF) to find the correspondences between the point clouds and the model for ICP registration matching.”
MethodRCPNet: Deep-Learning based Relative Camera Pose Estimation for UAVs
Yang, C., Liu, Y., & Zell, A. (2020) · International Conference on Unmanned Aircraft Systems (ICUAS) · doi:10.1109/icuas48674.2020.9214000 · Cited by 23
“Visual SLAM, as the mainstream of SLAM, estimates camera motion by extracting sparse keypoints (e.g., SIFT [1], SURF [2]) or directly from pixels as in dense SLAM systems [3], [4].”
BackgroundA SLAM Map Restoration Algorithm Based on Submaps and an Undirected Connected Graph
Zhan, Z., Jian, W., Li, Y., & Yang, Y. (2021) · IEEE Access · doi:10.1109/access.2021.3049864 · Cited by 18
Learning monocular 3D reconstruction of articulated categories from motion
Kokkinos, F., & Kokkinos, I. (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/cvpr46437.2021.00178 · Cited by 24
“Firstly, we draw inspiration from 3D vision which has traditionally relied on motion information for SFM [38, 11], SLAM [20, 28] or Non-Rigid SFM [39, 8, 7].”
BackgroundMethodBeyond Learning: Back to Geometric Essence of Visual Odometry via Fusion-Based Paradigm
Jin, Y., Wei, N., Fu, M., Liu, H., & Zhang, W. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3121483 · Cited by 17
A general approach for closed-loop registration in AR
Zheng, F., Schubert, R., & Weich, G. (2013) · IEEE Conference on Virtual Reality and 3D User Interfaces (VR) · doi:10.1109/vr.2013.6549358 · Cited by 16
Model Estimation and Selection towardsUnconstrained Real-Time Tracking and Mapping
Gauglitz, S., Sweeney, C., Ventura, J., Turk, M., & Höllerer, T. (2013) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2013.243 · Cited by 16
“Simultaneous Localization and Mapping (SLAM) systems such as [6], [8], [17], [27] can deal with environments of arbitrary geometry and any camera motion that induces parallax.”
BackgroundMethodInfluentialMulti-Spectral Visual Odometry without Explicit Stereo Matching
Dai, W., Zhang, Y., Sun, D., Hovakimyan, N., & Ping, L. (2019) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2019.00056 · Cited by 16
“Therefore, the photometric bundle adjustment [8], [9], [24] of multi-view stereo is used to achieve multi-spectral visual odometry.”
BackgroundMethodDENAO: Monocular Depth Estimation Network with Auxiliary Optical Flow
Chen, J., Yang, X. Y., Jia, Q., & Liao, C. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.2977021 · Cited by 18
BIT-VO: Visual Odometry at 300 FPS using Binary Features from the Focal Plane
Murai, R., Sajad Saeedi, G., & Kelly, P. H. J. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341151 · Cited by 22
Indoor Space Recognition using Deep Convolutional Neural Network: A Case Study at MIT CampusPreprint
Zhang, F., Duarte, F., Ma, R., Milioris, D., Lin, H. Z., & Ratti, C. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1610.02414 · Cited by 16
“It focusses on using various types of sensors to build a signal (ultrasonic, WiFi, geomagnetic, bluetooth) map for the target space [19,20].”
BackgroundFusion++: Volumetric Object-Level SLAMPreprint
McCormac, J., Clark, R., Bloesch, M., Davison, A. J., & Leutenegger, S. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1808.08378 · Cited by 36
“Unlike many dense reconstruction systems [24, 39, 43, 38, 3, 8] we make no attempt to keep a dense representation of the entire scene.”
BackgroundSelf-citationSpatial maps with working area limit line from images of crane's top-view camera
Wang, Y., Suzuki, H., & Ohtake, Y. (2022) · Automation in Construction · doi:10.1016/j.autcon.2022.104475 · Cited by 15
DSOL: A Fast Direct Sparse Odometry Scheme
Qu, C., Shivakumar, S. S., Miller, I. D., & Taylor, C. J. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981491 · Cited by 19
“DTAM [17] was the first system to demonstrate real-time camera tracking and dense scene reconstruction running on a GPU.”
BackgroundA review of visual SLAM with dynamic objects
Qin, Y. F., & Yu, H. (2023) · Industrial Robot the international journal of robotics research and application · doi:10.1108/ir-07-2023-0162 · Cited by 20
A Review of Crowdsourcing Update Methods for High-Definition Maps
Guo, Y., Zhou, J. Y., Li, X., Tang, Y., & Lv, Z. (2024) · ISPRS International Journal of Geo-Information · doi:10.3390/ijgi13030104 · Cited by 15
Implicit Event-RGBD Neural SLAM
Qu, D., Yan, C., Wang, D., Yin, J., Chen, Q., Xu, D., Zhang, Y., Zhao, B., & Li, X. (2024) · 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/cvpr52733.2024.01852 · Cited by 35
“Exploration in extreme environments [10] remains challenging for visual SLAM (vSLAM) systems due to the lack of visual features caused by motion blur and lighting variation in diverse environments [55, 63].”
BackgroundRobust Real-Time Visual Odometry for Stereo Endoscopy Using Dense Quadrifocal Tracking
Chang, P.-L., Handa, A., Davison, A. J., Stoyanov, D. V., & Edwards, P. (2014) · Lecture notes in computer science · doi:10.1007/978-3-319-07521-1_2 · Cited by 15
“To constrain tracking and prevent drift, we adopt frame-to-model tracking which is essentially the same concept as the keyframe strategy in visual SLAM systems [8, 12].”
BackgroundMethodSelf-citationA robust RGB-D SLAM system for 3D environment with planar surfaces
Su, P.-C., Shen, J., & Cheung, S.-C. S. (2013) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2013.6738057 · Cited by 15
“The camera pose estimation is further improved in [15].”
Backgroundσ-DVO: Sensor Noise Model Meets Dense Visual Odometry
Babu, B. P. W., Kim, S., Yan, Z., & Ren, L. (2016) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar.2016.11 · Cited by 18
“[18] presented Dense Tracking And Mapping (DTAM) which used a computationally expensive global optimization approach to perform pose estimation.”
MethodShah, U., Khawad, R., & Krishna, K. M. (2016) · Indian Conference on Computer Vision, Graphics and Image Processing · doi:10.1145/3009977.3010047 · Cited by 15
“ISBN 978-1-4503-4753-2/16/12. . . $15.00 DOI: http://dx.doi.org/10.1145/3009977.3010047 Keywords Micro Aerial Vehicles; Autonomous Navigation; Deep Learning”
BackgroundInfluentialCamera pose estimation for augmented reality in a small indoor dynamic scene
Frikha, R., Ejbali, R., & Zaied, M. (2017) · Journal of Electronic Imaging · doi:10.1117/1.jei.26.5.053029 · Cited by 15
3D Scene Mesh from CNN Depth Predictions and Sparse Monocular SLAM
Mukasa, T., Xu, J., & Stenger, B. (2017) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2017.112 · Cited by 18
“The first real-time method in this category was Dense Tracking and Mapping (DTAM) [22].”
BackgroundMethodInfluential
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.