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 301–400 of 1,004
Real-time panoramic tracking for event cameras
Reinbacher, C., Munda, G., & Pock, T. (2017) · IEEE International Conference on Computational Photography (ICCP) · doi:10.1109/iccphot.2017.7951488 · Cited by 78
“Our method for panoramic camera tracking is inspired by direct alignment approaches like DTAM [20], LSD-SLAM [7] and DSO [6].”
MethodReal-Time Motion Tracking for Mobile Augmented/Virtual Reality Using Adaptive Visual-Inertial Fusion
Fang, W., Zheng, L., Deng, H., & Zhang, H. (2017) · Sensors · doi:10.3390/s17051037 · Cited by 87
“Then, DTAM [19] and LSD-SLAM [20] were proposed; these tracking methods are based on the minimization of the photometric pixel values instead of the feature-based matching, thus possessing larger computational requirements.”
MethodSalientDSO: Bringing Attention to Direct Sparse Odometry
Liang, H.-J., Sanket, N. J., Fermuller, C. M., & Aloimonos, Y. (2019) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2019.2900980 · Cited by 75
“Such prior is necessary for reconstructing a dense world model [3, 21,25].”
Background3D point cloud map based vehicle localization using stereo camera
Xu, Y., John, V., Mita, S., Tehrani, H., Ishimaru, K., & Nishino, S. (2017) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2017.7995765 · Cited by 85
“Although the SLAM based vehicle localization algorithms make great improvement recently and the loop closure detection [19] techniques are used, it still suffers from the error accumulation and divergence.”
InfluentialLarge-scale outdoor 3D reconstruction on a mobile device
Schöps, T., Sattler, T., Häne, C., & Pollefeys, M. (2016) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2016.09.007 · Cited by 74
“[8, 10] employ regularization to deal with the latter.”
MethodInfluentialZero-Shot Depth Estimation From Light Field Using A Convolutional Neural Network
Peng, J., Xiong, Z., Wang, Y., Zhang, Y., & Liu, D. (2020) · IEEE Transactions on Computational Imaging · doi:10.1109/tci.2020.2967148 · Cited by 71
“Conventional depth from video methods such as SFM [40], OEB-SLAM [41] and DTAM [42], explore the scene geometry from images taken from different positions.”
MethodDirect-PoseNet: Absolute Pose Regression with Photometric Consistency
Chen, S., Wang, Z., & Prisacariu, V. A. (2021) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv53792.2021.00125 · Cited by 100
“Direct matching is a common approach in traditional SLAM and VO systems [10, 11, 13, 14, 36, 42].”
MethodPseudo RGB-D for Self-improving Monocular SLAM and Depth Prediction
Tiwari, L., Ji, P., Tran, Q.-H., Zhuang, B., Anand, S., & Chandraker, M. K. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58621-8_26 · Cited by 80
An Overview of Key SLAM Technologies for Underwater Scenes
Wang, X., Fan, X., Shi, P., Ni, J., & Zhou, Z. (2023) · Remote Sensing · doi:10.3390/rs15102496 · Cited by 80
“In contrast to the feature-based method, direct methods [77] such as SVO [32] and DSO [78] minimize the photometric error, and the function takes the form of subtracting the grayscale values of the pixels.”
MethodCurrent Status and Future Trends of Meter-Level Indoor Positioning Technology: A Review
Qi, L., Liu, Y., Yu, Y., Chen, L., & Chen, R. (2024) · Remote Sensing · doi:10.3390/rs16020398 · Cited by 85
Real-time surface light-field capture for augmentation of planar specular surfaces
Jachnik, J., Newcombe, R. A., & Davison, A. J. (2012) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar.2012.6402544 · Cited by 78
“A clear next step with this work is to seek to combine light-field capture with live dense reconstruction as in [13].”
BackgroundMethodInfluentialSelf-citationToward Domain Independence for Learning-Based Monocular Depth Estimation
Mancini, M., Costante, G., Valigi, P., Ciarfuglia, T. A., Delmerico, J., & Scaramuzza, D. (2017) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2017.2657002 · Cited by 81
“Despite the impressive results achieved by state-of-the-art approaches [14]–[16], the performance of their reconstruction routines drops during high-speed motion, as dense alignment becomes extremely challenging.”
BackgroundA Robust RGB-D SLAM System With Points and Lines for Low Texture Indoor Environments
Fu, Q., Yu, H., Lai, L., Wang, J., Xia, P., Sun, W. J., & Sun, M. (2019) · IEEE Sensors Journal · doi:10.1109/jsen.2019.2927405 · Cited by 75
“Early methods were computationally intensive, which was typically addressed using GPU parallelism, because they implement whole image alignment, such as LSD SLAM [7] and DTAM [19].”
MethodDirect Sparse OdometryPreprint
Engel, J. J., Koltun, V., & Cremers, D. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1607.02565 · Cited by 70
RelPose: Predicting Probabilistic Relative Rotation for Single Objects in the Wild
Zhang, J. Y., Ramanan, D., & Tulsiani, S. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19821-2_34 · Cited by 118
Gaussian Splatting: 3D Reconstruction and Novel View Synthesis: A Review
Dalal, A., Hagen, D., Robbersmyr, K. G., & Knausgård, K. M. (2024) · IEEE Access · doi:10.1109/access.2024.3408318 · Cited by 115
“Visual SLAM algorithms fall into two main categories: sparse methods, employing feature point matching (e.g., Parallel Tracking and Mapping [117], ORB-SLAM [118]), and dense methods, which utilize overall image brightness (e.g., DTAM [119], LSD-SLAM [120], DSO [121], SVO [122]).”
MethodTightly-Coupled Monocular Visual-Odometric SLAM Using Wheels and a MEMS Gyroscope
Quan, M., Piao, S., Tan, M., & Huang, S.-S. (2019) · IEEE Access · doi:10.1109/access.2019.2930201 · Cited by 74
Interest Point Detector and Feature Descriptor Survey
Krig, S. (2014) · Apress eBooks · doi:10.1007/978-1-4302-5930-5_6 · Cited by 80
Structure Aware SLAM Using Quadrics and Planes
Hosseinzadeh, M., Latif, Y., Pham, T. T., Sünderhauf, N., & Reid, I. D. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-20893-6_26 · Cited by 75
“s problem [13]. SLAM with cameras has also seen advancement in theory and good implementations that have lead to many real-time systems from sparse ([20],[6]) to semi-dense ([7], [8]) to fully dense ([23], [22], [24]). Recently, there has been a lot of interest in extending the capability of a pointbased representatio…”
BackgroundMethodPhotometric Mesh Optimization for Video-Aligned 3D Object Reconstruction
Lin, C.-H., Wang, O., Russell, B. C., Shechtman, E., Kim, V. G., Fisher, M., & Lucey, S. (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.2019.00106 · Cited by 79
“Classical multiview geometric methods, most notably structure frommotion (SfM) and SLAM, recover point clouds as the underlying 3D structure of RGB sequences, often with very high accuracy [10, 29].”
Line Flow Based Simultaneous Localization and Mapping
Wang, Q., Yan, Z., Wang, J., Xue, F., Ma, W., & Zha, H. (2021) · IEEE Transactions on Robotics · doi:10.1109/tro.2021.3061403 · Cited by 73
Czyża, S., Szuniewicz, K., Kowalczyk, K., Dumalski, A., Ogrodniczak, M., & Zieleniewicz, Ł. (2023) · Sensors · doi:10.3390/s23042092 · Cited by 80
“Among the basic visual odometry methods used in the field of drone positioning, we can mention solutions using algorithms such as SfM (Structure from Motion) [55,56], PTAM (Parallel Tracking and Mapping) [57] and DTAM (Dense Tracking and Mapping) [58].”
MethodA comprehensive overview of core modules in visual SLAM framework
Cai, D., Li, R., Hu, Z., Lu, J., Li, S., & Zhao, Y. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.127760 · Cited by 75
Using superpixels in monocular SLAM
Concha, A., & Civera, J. (2014) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2014.6906883 · Cited by 78
“Approaches like DTAM [6] present dense pixel-wise reconstructions that might be essential in certain applications –e.g. augmented reality– but come at the price of a large parallel computation in the GPU.”
BackgroundPhotometric Bundle Adjustment for Vision-Based SLAM
Alismail, H., Browning, B., & Lucey, S. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-54190-7_20 · Cited by 84
“Minimizing the photometric error, however, has been limited to frame–frame estimation (visual odometry), or as a tool for depth refinement independent of the parameters of motion [9].”
BackgroundResultInfluentialDeepVO: A Deep Learning approach for Monocular Visual Odometry
Mohanty, V., Agrawal, S., Datta, S., Ghosh, A., Sharma, V. D., & Chakravarty, D. (2016) · arXiv · doi:10.48550/arxiv.1611.06069 · Cited by 67
Real-time monocular dense mapping on aerial robots using visual-inertial fusion
Yang, Z., Gao, F., & Shen, S. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989529 · Cited by 71
“However, the convex optimization used in DTAM and REMODE requires a desktop GPU.”
MethodDeLS-3D: Deep Localization and Segmentation with a 3D Semantic Map
Wang, P., Yang, R., Cao, B., Xu, W., & Lin, Y. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00614 · Cited by 66
Jacobsen, E. L., Solberg, A., Golovina, O., & Teizer, J. (2021) · Construction Innovation · doi:10.1108/ci-06-2021-0113 · Cited by 77
Robust stereo visual odometry from monocular techniques
Persson, M., Piccini, T., Felsberg, M., & Mester, R. (2015) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2015.7225764 · Cited by 75
“Direct methods on the other hand operate on the whole images thus producing a dense or semi-dense motion field [11], [12], [13], [14].”
BackgroundMethodGN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization
von Stumberg, L., Wenzel, P., Khan, Q., & Cremers, D. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.2965031 · Cited by 81
“Popular direct visual methods are DTAM [16], LSD-SLAM [3], DSO [5], and PhotoBundle [4].”
MethodA Survey on Visual Navigation and Positioning for Autonomous UUVs
Qin, J., Li, M., Li, D., Zhong, J., & Yang, K. (2022) · Remote Sensing · doi:10.3390/rs14153794 · Cited by 65
“Many excellent algorithms have emerged in the tightly-coupled field, which can be roughly divided into two categories: filter-based methods and optimization-based methods according to the type of back-end optimization.”
MethodReal-Time Dynamic SLAM Algorithm Based on Deep Learning
Su, P., Luo, S., & Huang, X. (2022) · IEEE Access · doi:10.1109/access.2022.3199350 · Cited by 78
“Newcombe et al. [4] proposed the DTAM system, which realized the establishment of a dense map of every frame by the direct method.”
MethodDepth estimation and camera calibration of a focused plenoptic camera for visual odometry
Zeller, N., Quint, F., & Stilla, U. (2016) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2016.04.010 · Cited by 69
“As presented by Newcombe et al. (2011), direct methods can be used to estimate a dense depth map.”
MethodLoosely-Coupled Semi-Direct Monocular SLAM
Lee, S. H., & Civera, J. (2018) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2018.2889156 · Cited by 81
“Unlike most featurebased methods, direct methods are not limited to a sparse map and can maintain either a sparse [6], semi-dense [5], [17] or dense [18], [19] map in real-time.”
Vision guided robotic inspection for parts in manufacturing and remanufacturing industry
Khan, A., Mineo, C., Dobie, G., MacLeod, C., & Pierce, G. (2020) · Journal of remanufacturing · doi:10.1007/s13243-020-00091-x · Cited by 75
“In recent years, simultaneous localisation and mapping (SLAM) [7, 8, 10, 32] using only cameras has been actively discussed, as this sensor configuration is simple, operates at real-time and achieves drift-free performance inaccessible to Structure from Motion approaches.”
BackgroundOn the performance of metrics to predict quality in point cloud representations
Ebrahimi, T., & Alexiou, E. (2017) · SPIE Applications of Digital Image Processing · doi:10.1117/12.2275142 · Cited by 69
Loosely-Coupled Semi-Direct Monocular SLAM
Lee, S. H., & Civera, J. (2019) · Zaguan (University of Zaragoza Repository) · Cited by 63
Ground-Level Mapping and Navigating for Agriculture Based on IoT and Computer Vision
Zhao, W., Wang, X., Qi, B., & Runge, T. M. (2020) · IEEE Access · doi:10.1109/access.2020.3043662 · Cited by 65
Odometry-Vision-Based Ground Vehicle Motion Estimation With SE(2)-Constrained SE(3) Poses
Fan, Z., Tang, H., & Liu, Y. (2018) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2018.2831900 · Cited by 63
“DTAM [13] by Newcombe et al.”
Papachristos, C., Mascarich, F., Khattak, S., Dang, T., & Alexis, K. (2019) · Autonomous Robots · doi:10.1007/s10514-019-09864-1 · Cited by 67
“It is highlighted too that under such conditions standard direct visual odometry solutions (Newcombe et al. 2011; Engel et al. 2013) will be limited by constant brightness assumptions and the poor image illumination intensity levels, while more recent direct odometry approaches that deal with…”
BackgroundDense RGB-D-inertial SLAM with map deformations
Laidlow, T., Bloesch, M., Li, W., & Leutenegger, S. (2017) · 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2017.8206591 · Cited by 61
“Examples using solely monocular cameras range from the fully dense DTAM [14] to LSD-SLAM [15], which reconstructs a semi-dense representation.”
MethodDirect monocular odometry using points and lines
Yang, S., & Scherer, S. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989446 · Cited by 60
“Regularization term Gj makes the depth of pixels on one edge correlate with each other and is typically solved by an iterative alternating optimization through duality principles [12].”
BackgroundMethodInfluentialOn the Limits of Pseudo Ground Truth in Visual Camera Re-localisation
Brachmann, E., Humenberger, M., Rother, C., & Sattler, T. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00616 · Cited by 99
“KinectFusion [30, 48], an early incarnation of such a system, uses a truncated signed distance function (TSDF) to represent the scene.”
BackgroundOA-SLAM: Leveraging Objects for Camera Relocalization in Visual SLAM
Zins, M., Simon, G., & Berger, M. (2022) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar55827.2022.00090 · Cited by 80
“Many visual SLAM methods have been developed, turning RGB image sequences into sparse [7, 17, 23], semi-dense [8, 9], or dense [24] maps.”
MethodCanelhas, D. R., Stoyanov, T., & Lilienthal, A. J. (2013) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2013.6696880 · Cited by 72
“I. INTRODUCTION Work has been done in recent years to leverage the power of parallel hardware architectures to produce practical realtime solutions for detailed and accurate scene reconstruction (e.g. [1] ).”
BackgroundReal-Time Dense Monocular SLAM With Online Adapted Depth Prediction Network
Luo, H., Gao, Y., Wu, Y., Liao, C., Yang, X. Y., & Cheng, K.-T. T. (2018) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2018.2859034 · Cited by 67
“There exists a vast body of literature on SLAM [2], [3], [14]–[17], depth estimation [12], [18]–[21], and 3D reconstruction [22], [23].”
BackgroundMethodMonocular Visual Odometry Based on Depth and Optical Flow Using Deep Learning
Ban, X., Wang, H., Chen, T., Wang, Y., & Xiao, Y. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.3024011 · Cited by 67
“For a typical SLAM algorithm with the VO of direct methods, DTAM [25] takes advantage of a dense depth map for each key-frame to minimize the global energy function by aligning the entire image.”
MethodCodeMapping: Real-Time Dense Mapping for Sparse SLAM using Compact Scene Representations
Matsuki, H., Scona, R., Czarnowski, J., & Davison, A. J. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3097258 · Cited by 74
“Several methods have attempted to perform dense mapping and camera localization simultaneously [6], [7],”
Self-citationDT-SLAM: Deferred Triangulation for Robust SLAM
Herrera, D., Kim, K., Kannala, J., Pulli, K., & Heikkilä, J. (2014) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2014.49 · Cited by 65
“Modeling an environment, and tracking the camera motion with respect to the environment, is a key component of many mobile vision and augmented reality (AR) applications, and it has been referred to as SLAM (simultaneous localization and mapping [7, 19, 13, 27]) and TAM (tracking and mapping [14, 20]).”
BackgroundComparative analysis of ROS-based monocular SLAM methods for indoor navigation
Buyval, A., Afanasyev, I. M., & Magid, E. (2017) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2268809 · Cited by 59
“…decade a number of monocular SLAM-related algorithms have appeared: Parallel Tracking and Mapping (PTAM,[3]), Dense Tracking and Mapping (DTAM, [4]), Large-Scale Direct Monocular SLAM (LSD-SLAM,[1]), REgularized MOnocular Depth Estimation (REMODE, [5]), ORB-SLAM [6], Dense Piecewise Planar…”
NID-SLAM: Robust Monocular SLAM Using Normalised Information Distance
Pascoe, G., Maddern, W., Tanner, M., Piniés, P., & Newman, P. M. (2017) · 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.2017.158 · Cited by 68
“Recent methods that directly minimise the photometric error between image and map are designed to address these limitations [32], providing increased accuracy, dense reconstructions and some robustness to viewpoint change and blur [7, 24].”
MethodLaidlow, T., Czarnowski, J., & Leutenegger, S. (2019) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2019.8793527 · Cited by 74
“Dense reconstruction systems, such as DTAM [4] and REMODE [5], typically create dense scene representations by optimising for the depth values that minimise the photometric error over several frames.”
BackgroundMethodIntegrating depth and color cues for dense multi-resolution scene mapping using RGB-D cameras
Stückler, J., & Behnke, S. (2012) · IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI) · doi:10.1109/mfi.2012.6343050 · Cited by 62
“Current work on SfM in computer vision also includes real-time dense surface reconstruction from monocular videos [9], [10].”
BackgroundTian, F., Gao, Y., Fang, Z., Fang, Y., Gu, J., Fujita, H., & Hwang, J. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3080928 · Cited by 61
Stewart, A. D., & Newman, P. M. (2012) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2012.6224750 · Cited by 60
“Perhaps the work for which the guiding intuition is most similar to ours is the localisation component of DTAM (Direct Tracking and Mapping) [7], in which the image feed is used in parallel to both estimate the 3D structure of the local scene and localise the camera within the scene.”
BackgroundRobust 3D Tracking with Descriptor Fields
Crivellaro, A., & Lepetit, V. (2014) · 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.2014.436 · Cited by 59
“Dense alignment is attractive because it globally exploits most of the image information, even when local image features such as interest points or edges are ambiguous.”
BackgroundDeep Learning for Visual SLAM in Transportation Robotics: A review
Duan, C., Junginger, S., Huang, J., Jin, K., & Thurow, K. (2019) · Transportation Safety and Environment · doi:10.1093/tse/tdz019 · Cited by 63
Semi-dense 3D Reconstruction with a Stereo Event Camera
Zhou, Y. B., Gallego, G., Rebecq, H., Kneip, L., Li, H., & Scaramuzza, D. (2018) · Zurich Open Repository and Archive (University of Zurich) · doi:10.5167/uzh-176054 · Cited by 58
3D High Dynamic Range dense visual SLAM and its application to real-time object re-lighting
Meilland, M., Barat, C., & Comport, A. I. (2013) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar.2013.6671774 · Cited by 67
“Approaches in dense localisation and mapping [3, 27, 33, 41, 25], however, have shown that dense 3D tracking can be performed in real-time using the fullimage.”
BackgroundMulti-Channel Generalized-ICP: A robust framework for multi-channel scan registration
Servos, J., & Waslander, S. L. (2016) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2016.10.016 · Cited by 59
“Several dense camera SLAM methods have been proposed including Dense Tracking and Mapping (DTAM) [16], created by Newcombe et al.”
MethodInfluentialHandheld Guides in Inspection Tasks: Augmented Reality versus Picture
Polvi, J., Taketomi, T., Moteki, A., Yoshitake, T., Fukuoka, T., Yamamoto, G., Sandor, C., & Kato, H. (2017) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2017.2709746 · Cited by 58
Estimating 3-D Point-of-Regard in a Real Environment Using a Head-Mounted Eye-Tracking System
Takemura, K., Takahashi, K., Takamatsu, J., & Ogasawara, T. (2014) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2014.2318324 · Cited by 56
“A dense map, typified by DTAM [18] and KinectFusion [19], is required, and this will form part of our future research.”
Visual industrial inspection using aerial robots
Omari, S., Gohl, P., Burri, M., Achtelik, M., & Siegwart, R. (2014) · International Conference on Applied Robotics for the Power Industry (CARPI) · doi:10.1109/carpi.2014.7030056 · Cited by 61
“Lastly, we evaluate the navigation system in a set of experiments.”
MethodThe Semantic Paintbrush: Interactive 3D Mapping and Recognition in Large Outdoor Spaces
Miksik, O., Vineet, V., Lidegaard, M., Prasaath, R., Nießner, M., Golodetz, S., Hicks, S. L., Pérez, P., Izadi, S., & Torr, P. H. S. (2015) · Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems · doi:10.1145/2702123.2702222 · Cited by 57
Polarimetric Dense Monocular SLAM
Yang, L., Tan, F., Li, A., Cui, Z., Furukawa, Y., & Tan, P. (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.2018.00406 · Cited by 61
“Similar to [28], we introduce an auxiliary variable a to simplify this problem, which turns the original energy function into,”
MethodGuided Feature Selection for Deep Visual Odometry
Xue, F., Wang, Q., Wang, X., Dong, W., Wang, J., & Zha, H. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-20876-9_19 · Cited by 61
ORB-TEDM: An RGB-D SLAM Approach Fusing ORB Triangulation Estimates and Depth Measurements
Yuan, J., Zhu, S., Tang, K., & Sun, Q. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3154800 · Cited by 65
Stereo vision and IMU based real-time ego-motion and depth image computation on a handheld device
Schmid, K., & Hirschmüller, H. (2013) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2013.6631242 · Cited by 64
“Modern approaches aim at matching all pixels, which require graphic card implementations due to the high computational load [4] and partly rely on active sensors [5], [6] that do not work outdoors due to sunlight.”
BackgroundEdge SLAM: Edge Points Based Monocular Visual SLAM
Maity, S., Saha, A., & Bhowmick, B. (2017) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2017.284 · Cited by 63
“Direct SLAM: Direct methods [26] gains popularity for its semi dense map creation.”
MethodDeep Keypoint-Based Camera Pose Estimation with Geometric Constraints
Jau, Y.-Y., Zhu, R., Su, H., & Chandraker, M. K. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341229 · Cited by 72
“This issue motivated the direct methods [12], [19], which maximizes photometric consistency over all pixel pairs.”
MethodFutureMapping: The Computational Structure of Spatial AI SystemsPreprint
Davison, A. J. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1803.11288 · Cited by 103
“Mainstream geometrical computer vision started to take advantage of parallel processing in the form of GPUs nearly 10 years ago (e.g. [45]), and in Spatial AI this led to break-throughs in dense SLAM [42, 41].”
BackgroundMethodInfluentialSelf-citationROSEFusion: random optimization for online dense reconstruction under fast camera motion
Zhang, J., Zhu, C., Zheng, L., & Xu, K. (2021) · ACM Transactions on Graphics · doi:10.1145/3450626.3459676 · Cited by 463
3D Reconstruction with Spatial Memory
Wang, H., & Agapito, L. (2025) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv66043.2025.00013 · Cited by 356
“…solutions evolve into various sub-fields, including keypoint detection and matching [10, 46, 47, 60], Structure-from-Motion (SfM) [2, 19, 64, 68, 75, 84, 85], Bundle Adjustment (BA) [3, 77, 86], Multi-View Stereo (MVS) [29, 30, 65], Simultaneous Localization and Mapping (SLAM) [21, 41, 54], etc.”
Incremental Visual-Inertial 3D Mesh Generation with Structural Regularities
Rosinol, A., Sattler, T., Pollefeys, M., & Carlone, L. (2019) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2019.8794456 · Cited by 68
Active monocular localization: Towards autonomous monocular exploration for multirotor MAVs
Mostegel, C., Wendel, A., & Bischof, H. (2014) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2014.6907417 · Cited by 55
“Current monocular SLAM approaches [8], [9], [10], [11] strive to passively determine the current camera pose and map the environment.”
BackgroundPartial Person Re-identification with Part-Part Correspondence Learning
He, T., Shen, X., Huang, J., Chen, Z., & Hua, X. (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/cvpr46437.2021.00899 · Cited by 61
“Learning correspondence between two pairs is a central topic in computer vision, and has been widely applied in image stitching [49], image restoration [5], object recognition [32], tracking [37, 7], medical imaging [1], etc.”
Zeng, T., Tang, F., Ji, D., & Si, B. (2020) · Neural Networks · doi:10.1016/j.neunet.2020.02.023 · Cited by 57
Zheng, Z., Lin, S., & Yang, C. (2024) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2024.3363744 · Cited by 65
“In the 21st century, visual and visual-inertial SLAM (VI-SLAM) gradually matured, giving rise to excellent open-source algorithms like DTAM [1], LSD-SLAM [2], ORB-SLAM3 [3], MSCKF [4], VINS-mono [5], etc.”
BackgroundRoDyn-SLAM: Robust Dynamic Dense RGB-D SLAM With Neural Radiance Fields
Jiang, H., Xu, Y., Li, K., Feng, J., & Zhang, L. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3427554 · Cited by 91
“Similar to most modern SLAM systems [32], [33], our system comprises two distinct processes: the tracking process as the frontend and the mapping process as the backend, combined with keyframe management { F k } Mk =1 and neural implicit map f θ .”
Direct Visual Localisation and Calibration for Road Vehicles in Changing City Environments
Pascoe, G., Maddern, W., & Newman, P. M. (2015) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2015.23 · Cited by 60
“The method in [21] builds a dense scene representation using all pixels from the input camera images, and demonstrates robustness to significant viewpoint changes and motion blur, but has only been tested in small indoor environments.”
MethodRobust visual-inertial localization with weak GPS priors for repetitive UAV flights
Surber, J., Teixeira, L., & Chli, M. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989745 · Cited by 55
“using denser and computationally more demanding approaches [18].”
MethodA Benchmark for Visual-Inertial Odometry Systems Employing Onboard Illumination
Kasper, M., McGuire, S., & Heckman, C. (2019) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros40897.2019.8968554 · Cited by 52
“This is particularly problematic for direct methods, which work on image intensities [8], [15], [4].”
BackgroundComparison of modern open-source Visual SLAM approaches
Sharafutdinov, D., Griguletskii, M., Kopanev, P., Kurenkov, M., Ferrer, G., Бурков, А. В., Gonnochenko, A., & Tsetserukou, D. (2023) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-023-01812-7 · Cited by 63
Local-to-Global Registration for Bundle-Adjusting Neural Radiance Fields
Chen, Y., Chen, X., Wang, X., Zhang, Q., Guo, Y., Shan, Y., & Wang, F. (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/cvpr52729.2023.00799 · Cited by 58
“S f M and SLAM.”
BackgroundInfluentialDense Visual SLAM with Probabilistic Surfel Map
Yan, Z., Ye, M., & Ren, L. (2017) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2017.2734458 · Cited by 58
“Newcombe et al. [23] proposed the Dense Tracking and Mapping (DTAM) which directly created a global dense mesh of the environment during the visual SLAM and also utilized the generated dense mesh for front-end tracking.”
BackgroundPothole 3D Reconstruction With a Novel Imaging System and Structure From Motion Techniques
Ahmed, A., Ashfaque, M., Haq, M. U. U., Mathavan, S., Kamal, K., & Rahman, M. M. (2021) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2021.3054026 · Cited by 52
“Without any external knowledge about the environment, the resulting reconstruction can only be defined for a scale [32].”
BackgroundResultSemi-direct EKF-based monocular visual-inertial odometry
Tanskanen, P., Nägeli, T., Pollefeys, M., & Hilliges, O. (2015) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2015.7354242 · Cited by 61
“volume[16] or on computing constrained scene flow [14].”
BackgroundResultSajad Saeedi, G., Spink, T., Gorgovan, C., Webb, A. M., Clarkson, J., Tomusk, E., Debrunner, T., Kaszyk, K., Marugán, P. G. D. A., Rodchenko, A., et al. (2018) · Proceedings of the IEEE · doi:10.1109/jproc.2018.2856739 · Cited by 57
“Modern dense volumetric methods based on signed distance functions such as DTAM [6] or explicit point clouds, such as ElasticFusion [33], are able to recover high quality geometric information in real time.”
InfluentialSelf-citationRobust Visual-Lidar Simultaneous Localization and Mapping System for UAV
Qian, J., Chen, K., Chen, Q., Yang, Y., Zhang, J., & Chen, S. (2021) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2021.3099166 · Cited by 55
“DTAM [19] approach is one of the first real-time direct visual SLAM implementations, but it relies heavily on the GPU to make this happen.”
Near-Field Perception for Low-Speed Vehicle Automation Using Surround-View Fisheye Cameras
Eising, C., Horgan, J., & Yogamani, S. K. (2021) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2021.3127646 · Cited by 78
“Dense Tracking and Mapping (DTAM) [98] and”
Camera, LiDAR and Multi-modal SLAM Systems for Autonomous Ground Vehicles: a Survey
Chghaf, M., Rodríguez, S., & Elouardi, A. (2022) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-022-01582-8 · Cited by 62
“In 2011 Newcombe et al. proposed DTAM [88] which is one of the first direct methods, it computes dense depth map by minimizing a global spatially regularized energy function.”
BackgroundMethodInfluentialMulti-level mapping: Real-time dense monocular SLAM
Greene, W. N., Ok, K., Lommel, P., & Roy, N. (2016) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2016.7487213 · Cited by 55
“We have presented a monocular SLAM algorithm capable of operating in real-time without GPU acceleration that significantly improves upon the state-of-the-art in terms of tracking accuracy, reconstruction accuracy, and reconstruction density.”
BackgroundMethodInfluentialCopyMe3D: Scanning and Printing Persons in 3D
Sturm, J., Bylow, E., Kahl, F., & Cremers, D. (2013) · Lecture notes in computer science · doi:10.1007/978-3-642-40602-7_43 · Cited by 64
“More recent examples include [11, 4, 15] using monocular and stereo cameras, respectively.”
BackgroundEstimating Drivable Collision-Free Space from Monocular Video
Yao, J., Ramalingam, S., Taguchi, Y., Miki, Y., & Urtasun, R. (2015) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2015.62 · Cited by 58
“Monocular videos have been used to develop SLAM algorithms [28, 11, 22, 29, 27].”
MethodA Memory-Efficient KinectFusion Using Octree
Zeng, M., Zhao, F., Zheng, J., & Liu, X. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-34263-9_30 · Cited by 56
“Recently, some studies used RGB sequence to reconstruct the dense surface [12,13,10] of a small scene (about 1m× 1m× 1m) in real time.”
BackgroundSemi-dense SLAM on an FPGA SoC
Boikos, K., & Bouganis, C.-S. (2016) · International Conference on Field Programmable Logic and Applications (FPL) · doi:10.1109/fpl.2016.7577365 · Cited by 57
“As a result, they require at the very least a modern multicore desktop CPU to run in real time, and often high-end GPUs for acceleration as well [5], [6].”
BackgroundTaking a Deeper Look at the Inverse Compositional Algorithm
Lv, Z., Dellaert, F., Rehg, J. M., & Geiger, A. (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.2019.00471 · Cited by 56
“Classical approaches address this problem by exploiting more robust objective functions [34, 39], heuristically chosen patch-[49] or gradient-based [31] features, and photometric calibration as a pre-processing step [18].”
Efficient compositional approaches for real-time robust direct visual odometry from RGB-D data
Klose, S., Heise, P., & Knoll, A. (2013) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2013.6696487 · Cited by 55
“[16], where a forward-compositional formulation for the motion of a single camera is used.”
BackgroundXu, Z., Liu, Y., Shi, X., Wang, Y., & Zheng, Y. (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/cvpr42600.2020.00602 · Cited by 51
“Multi-view stereo (MVS) is a hot research area in computer vision, which offers a cheap and convenient way to capture the 3D geometry of scenes and objects, and serves as a main ingredient for many CV applications, e.g. [40, 26, 25, 24, 8].”
Background
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.