The DOI matches its Crossref record; 29 of the 29 references it cites were found in a registry or an index.
DeepVO: Towards end-to-end visual odometry with deep Recurrent Convolutional Neural Networks
Open accessAbstract
This paper studies monocular visual odometry (VO) problem. Most of existing VO algorithms are developed under a standard pipeline including feature extraction, feature matching, motion estimation, local optimisation, etc. Although some of them have demonstrated superior performance, they usually need to be carefully designed and specifically fine-tuned to work well in different environments. Some prior knowledge is also required to recover an absolute scale for monocular VO. This paper presents a novel end-to-end framework for monocular VO by using deep Recurrent Convolutional Neural Networks (RCNNs). Since it is trained and deployed in an end-to-end manner, it infers poses directly from a sequence of raw RGB images (videos) without adopting any module in the conventional VO pipeline. Based on the RCNNs, it not only automatically learns effective feature representation for the VO problem through Convolutional Neural Networks, but also implicitly models sequential dynamics and relations using deep Recurrent Neural Networks. Extensive experiments on the KITTI VO dataset show competitive performance to state-of-the-art methods, verifying that the end-to-end Deep Learning technique can be a viable complement to the traditional VO systems.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“More surprisingly, they show that the welltrained networks enable scale-consistent predictions over long videos, while the accuracy is still inferior to traditional methods because of ignoring geometric information.”
From DF-VO: What Should Be Learnt for Visual Odometry? (2021) · cites this paper as Wang et al. 2017
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“On the opposing side, end-to-end approaches can estimate the 6 degrees of freedom (DoF) poses directly from a sequence of RGB images, that is, it does not need engineering effort to design the modules in the traditional VO pipeline [4].”
From Transformer-Based Model for Monocular Visual Odometry: A Video Understanding Approach (2023) · cites this paper as [4]
The cited source says what the claim attributes to it.
In the citing paper
“This problem is called monocular visual odometry and often relies on geometric approaches that require considerable engineering effort for a specific scenario.”
From Transformer-Based Model for Monocular Visual Odometry: A Video Understanding Approach (2023) · cites this paper as [4]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“the amount of crater detections that ODN-detector produces is robust to disturbances or modifications of the image quality due to Gaussian noise, shot noise, brightness levels, as well as different illumination conditions.”
From Optical navigation for Lunar landing based on Convolutional Neural Network crater detector (2022) · cites this paper as [16]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 1×
“There have been some popular and powerful DNN architectures, such as VGGNet [22] and GoogLeNet [23], developed for computer vision tasks, producing remarkable performance.”
MethodGoing deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562
Are we ready for autonomous driving? The KITTI vision benchmark suite
Geiger, A., Lenz, P., & Urtasun, R. (2012) · 2012 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2012.6248074 · Cited by 14,897 · cited 2×
“In this section, we evaluate the proposed end-to-end monocular VO approach on the well-known KITTI VO/SLAM benchmark [3].”
BackgroundMethod(2016) · Cited by 11,679
ORB-SLAM: A Versatile and Accurate Monocular SLAM System
Mur-Artal, R., Montiel, J. M. M., & Tardos, J. D. (2015) · IEEE Transactions on Robotics · doi:10.1109/tro.2015.2463671 · Cited by 6,778 · cited 1×
“Examples include keyframe base PTAM [9] and ORB-SLAM [10].”
BackgroundParallel 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 · cited 1×
“Examples include keyframe base PTAM [9] and ORB-SLAM [10].”
BackgroundFlownet: Learning optical flow with convolutional networks
Dosovitskiy, A., Fischer, P., Ilg, E., Hausser, P., Hazirbas, C., Golkov, V., Smagt, P. V. D., Cremers, D., & Brox, T. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.316 · Cited by 4,313 · cited 2×
“In order to reduce both the training time and data required to converge, the CNN is based on a pre-trained FlowNet model [24].”
MethodMonoSLAM: Real-time single camera SLAM
Davison, A. J., Reid, I. D., Molton, N. D., & Stasse, O. (2007) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2007.1049 · Cited by 3,959 · cited 1×
“To mitigate this problem, visual Simultaneous Localisation and Mapping (SLAM) or Structure from Motion (SfM) can be adopted to maintain a feature map for drift correction along with pose estimation [8].”
MethodORB-SLAM: A Versatile and Accurate Monocular SLAM System
Mur-Artal, R., Montiel, J. M. M., & Tardos, J. D. (2015) · Zaguan (Universidad de Zaragoza) · Cited by 3,153
Multiple view geometry in computer vision
Opower, H. (2002) · Optics and Lasers in Engineering · doi:10.1016/s0143-8166(01)00145-2 · Cited by 3,044
Research works printed but not linked: 1
- zaremba, 2014, Learning to execute
Cited by
Sorted by citations
Items 401–500 of 969
Deep Learned Ground Penetrating Radar Subsurface Features for Robot Localization
Wickramanayake, S., Thiyagarajan, K., & Kodagoda, S. (2022) · Proceedings of IEEE Sensors/Proceedings of IEEE Sensors ... · doi:10.1109/sensors52175.2022.9967350 · Cited by 11
“Research has been done to develop deep learning techniques that can estimate the odometry of a robot using images captured by RGB cameras without the need for wheel encoders [14], [15].”
BackgroundStreet Floor Segmentation for a Wheeled Mobile Robot
Hyun, J., Woo, S., & Kim, E. (2022) · IEEE Access · doi:10.1109/access.2022.3227203 · Cited by 11
“For example, the classical loop closing is replaced with the learning-based image matching [3], [4], and the classical visual odometry (VO) is being combined with the learning-based VO [5], [6], [7].”
MethodCbwLoss: Constrained Bidirectional Weighted Loss for Self-Supervised Learning of Depth and Pose
Wang, F., Cheng, J. S., & Liu, P. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3250744 · Cited by 9
“Based on the method of [35], the uncertainty of the camera pose has been considered and modeled using its covariance to correct the drift and bound the uncertainty [36].”
BackgroundMethodCAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View DepthPreprint
Fácil, J. M., Ummenhofer, B., Zhou, H., Montesano, L., Brox, T., & Civera, J. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1904.02028 · Cited by 10
“…[47], from single [6] and multiple views [33], using end-to-end networks [21] and fusing with multi-view geometry [8], completing depth maps [45, 42], and estimating geolocation [43, 19], relative motion [38], visual odometry [40, 41], and simultaneous localization and mapping (SLAM) [35, 2, 46].”
BackgroundDeep Learning based Monocular Depth Prediction: Datasets, Methods and ApplicationsPreprint
Li, Q., Zhu, J., Liu, J., Cao, R., Li, Q., Jia, S., & Qiu, G. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.04123 · Cited by 12
“[184] utilize a virtual stereo view to predict depth.”
MethodEnd-to-End Deep Learning Framework for Real-Time Inertial Attitude Estimation using 6DoF IMUPreprint
Golroudbari, A. A., & Sabour, M. H. (2023) · doi:10.22541/au.167628608.89589390/v1 · Cited by 9
Sequential learning navigation method and general correction model for Autonomous Underwater Vehicle
Zhang, X., He, B., Gao, S., Zhou, L., & Huang, R. (2023) · Ocean Engineering · doi:10.1016/j.oceaneng.2023.114347 · Cited by 9
Navigation Sensor Data Reliability Model-Based on Self-Evaluation and Mutual Evaluation
Li, W., Zhang, Z., Liang, Y. C., Shen, F., Gao, W. H., & Xu, D. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3304753 · Cited by 8
Robust Embedded Autonomous Driving Positioning System Fusing LiDAR and Inertial Sensors
He, Z., Xue, B., Hu, X., Shen, Z., Zeng, X., & Liu, M. (2023) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3626098 · Cited by 8
“Many Convolutional Neural Networks (CNNs) architectures [61], [51], [59] have put their efect into LiDAR classiication and segmentation problem, but CNN-based 6 degrees of freedom (6-DoF) pose estimation remains minority due to its inaccurate depth prediction and scale drift.”
BackgroundWei, W., Yang, P., Li, J., Liu, X., & Zhou, Y. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3404924 · Cited by 9
“On the other hand, many learning-based MVOs are supervised models [5], [6], which require large-scale accurately annotated ground-truth datasets.”
BackgroundMethodResearch on Low-Light Environment SLAM System via Retinex Enhancement and Deep Feature Optimization
Jiang, K., Zhu, C., & Yang, J. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3581532 · Cited by 9
“The first is end-to-end methods, such as DeepVO [13] and TartanVO [14], which directly predict pose parameters from continuous image frames.”
Ebrahimi, A., Mosavi, M. R., & Ayatollahi, A. (2025) · GPS Solutions · doi:10.1007/s10291-025-01912-x · Cited by 8
Knowledge Distillation in Object Detection: A Survey from CNN to Transformer
Shehzadi, T., Noor, R., Ifza, I., Liwicki, M., Stricker, D., & Afzal, M. Z. (2026) · Sensors · doi:10.3390/s26010292 · Cited by 10
DecoupledPoseNet: Cascade Decoupled Pose Learning for Unsupervised Camera Ego-Motion Estimation
Zhou, W., Zhang, H., Yan, Z., Wang, W., & Lin, L. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3144958 · Cited by 8
Zhang, Q., Hao, L., Li, H., Ren, Z., & Cheng, J. S. (2022) · International Symposium on Autonomous Systems (ISAS) · doi:10.1109/isas55863.2022.9757267 · Cited by 8
General spiking neural network framework for the learning trajectory from a noisy mmWave radar
Liu, X., Yan, M., Deng, L., Wu, Y., Han, D. G., Li, G., Ye, X., & Fan, D. (2022) · Neuromorphic Computing and Engineering · doi:10.1088/2634-4386/ac889b · Cited by 8
Robust visual odometry using sparse optical flow network
Liu, Q., & Chen, B. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.105471 · Cited by 10
AFT-VO: Asynchronous Fusion Transformers for Multi-View Visual Odometry Estimation
Kaygusuz, N., Méndez, O., & Bowden, R. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981835 · Cited by 8
“We use the PyTorch implementation [12] of DeepVO.”
BackgroundMethodInfluentialVision-aided intelligent and adaptive vehicle pose estimation during GNSS outages
Xu, Q., Tian, Y., Ruan, G., & Chang, B. (2023) · Measurement Science and Technology · doi:10.1088/1361-6501/ad14e4 · Cited by 8
“The translation– motion perception module refers to the DeepVO [13], and the rotation–motion perception module refers to our previous work on using deep learning for heading angle estimation [22].”
Learning‐based monocular visual‐inertial odometry with SE2(3) $S{E}_{2}(3)$‐EKF
Guo, C., Hu, J., & Luo, Y. (2024) · Journal of Field Robotics · doi:10.1002/rob.22349 · Cited by 8
“…devote to deal with the drawbacks of VO and VIO through learning ‐ based methodologies, and propose a series of learning ‐ based VO algorithms (Li et al., 2018; Wang et al., 2017, 2018; Yin & Shi, 2018) and VIO algorithms (Clark, Wang, Wen, et al., 2017; Han et al., 2019; Li & Waslander, 2020).”
InfluentialVONAS: Network Design in Visual Odometry using Neural Architecture Search
Cai, X., Zhang, L., Li, C., Li, G., & Li, T. H. (2020) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3394171.3413866 · Cited by 8
“Methods Paras(M) GMACs Avg terr Avg rerr DeepVO [21] 81.”
BackgroundMethodInfluentialPose Estimation Utilizing a Gated Recurrent Unit Network for Visual Localization
Kim, S., Kim, I.-H., Vecchietti, L. F., & Har, D. (2020) · Applied Sciences · doi:10.3390/app10248876 · Cited by 8
“In Table 1, the comparison with the Deep VO [23], that is one of the learning-based VO methods, is presented.”
MethodLearning Optical Flow with R-CNN for Visual Odometry
Huang, Y., Zhao, B., Gao, C., & Hu, X. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9562074 · Cited by 8
“A typical RPR VO was DeepVO network proposed by Wang [6] et al. which used the stacked consecutive images as input to estimate relative camera pose.”
BackgroundMethodResultFeature-aided Bundle Adjustment Learning Framework for Self-supervised Monocular Visual Odometry
Mai, W., & Watanabe, Y. (2021) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros51168.2021.9636756 · Cited by 8
“Su-pervised learning methods suggest the use of deep convolutional neural networks (CNNs) and recurrent neural networks (RNNs) [45], [46], [50] to model spatial-temporal information in data for pose estimation, but the large computational cost restricts the throughput, and the time-consuming ground…”
BackgroundTowards Interpretable Camera and LiDAR Data Fusion for Autonomous Ground Vehicles Localisation
Tibebu, H., Silva, D., Artaud, C., Pina, R., & Shi, X. (2022) · Sensors · doi:10.3390/s22208021 · Cited by 8
“[21] argued that, while CNNs are suitable for capturing useful learning features, they are not satisfiable enough to extract motion dynamics.”
BackgroundOptical encoder neural network: a CNN-based optical encoder for robot localization
Patruno, C., Renò, V., Nitti, M., Pernisco, G., & Mosca, N. (2022) · Optical Engineering · doi:10.1117/1.oe.62.4.041402 · Cited by 8
“18 These networks allow for both learning an effective feature representation for VO applications and modeling sequential relations and dynamics implicitly.”
BackgroundFrame-to-Frame Visual Odometry Estimation Network With Error Relaxation Method
Hwang, S., Cho, M., Ban, Y., & Lee, K. (2022) · IEEE Access · doi:10.1109/access.2022.3214823 · Cited by 9
“Wang et al. [34] suggested an RNN, called DeepVO, to estimate accurate visual odometry using prior knowledge.”
MethodA Survey on Visual SLAM Algorithms Compatible for 3D Space Reconstruction and Navigation
Khoyani, A., & Amini, M. (2023) · International Conference on Consumer Electronics · doi:10.1109/icce56470.2023.10043432 · Cited by 8
“In [23]–[25], a visual odometry was implemented while in [26], and [27]a local and global mapping were developed.”
MethodMOFT: Monocular odometry based on deep depth and careful feature selection and tracking
Koledić, K., Cvišić, I., Marković, I., & Petrovic, I. K. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160588 · Cited by 10
“Endto-end learning has been used in both supervised [6] and self-supervised formulations [7], [8], where networks usually jointly learn depth maps and ego-motion.”
MethodDVONet: Unsupervised Monocular Depth Estimation and Visual Odometry
Li, X., Hou, Y., Wu, Q., Wang, P., & Li, W. (2019) · IEEE International Conference on Visual Communications and Image Processing (VCIP) · doi:10.1109/vcip47243.2019.8965952 · Cited by 8
“[3] proposed an end-toend supervised monocular visual odometry (VO) system called “DeepVO”, which achieved comparable results to traditional VO methods on the KITTI dataset [4].”
MethodResultAerial Geo-Localisation for MAVs using PoseNet
Cabrera-Ponce, A. A., & Martínez-Carranza, J. (2019) · Workshop on Research, Education and Development of Unmanned Aerial Systems (RED-UAS) · doi:10.1109/reduas47371.2019.8999713 · Cited by 8
“Usually, the VO has used with monocular cameras and deep learning, allowing it to estimate the motion trajectory and infer poses directly from the camera using dataset and architectures like AlexNet [8], RCNNs [9] and unsupervised deep learning schemes based on VGG architecture [10].”
MethodMVP: Unified Motion and Visual Self-Supervised Learning for Large-Scale Robotic NavigationPreprint
Chancán, M., & Milford, M. J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.00667 · Cited by 8
“multi-sensor fusion techniques combining vision [24]–[30], inertial sensors [31]–[33], LiDAR [34]–[36], and radar [37] have been proposed to further improve both the localization accuracy and the real-time performance, with a number of recent deep-learning-based methods that can match the precision of those traditiona…”
MethodExploring Self-Attention for Visual OdometryPreprint
Damirchi, H., Khorrambakht, R., & Taghirad, H. D. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.08634 · Cited by 8
“Moreover, due to the similarity of the base structure of our network to [16], we are able to compare our attentionbased VO network directly against this SOTA attention-less network.”
InfluentialConwell, C., Mayo, D., Buice, M. A., Katz, B., Alvarez, G. A., & Barbu, A. (2021) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2021.06.18.448431 · Cited by 11
“for visual odometry [92]) are promising candidates for inclusion in future benchmarks, and our neural encoding & representational similarity metrics are just two of many variants.”
BackgroundGolroudbari, A. A., & Sabour, M. H. (2023) · doi:10.22541/au.168664884.43899660/v2 · Cited by 9
NAIR: An Efficient Distributed Deep Learning Architecture for Resource Constrained IoT System
Xiao, Y., Zhang, D., Wang, Y., Dai, X., Huang, Z., Zhang, W., Yang, Y., Anjum, A., & Qin, F. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3375924 · Cited by 7
“I N RECENT years, deep learning rapidly developed in various fields, including computer vision [1], [2], natural language processing [3], [4], and visual odometry [5], [6].”
BackgroundMBRVO: A Blur Robust Visual Odometry Based on Motion Blurred Artifact Prior
Zhang, J., Li, J., Li, J., Sun, Y., Liu, X., Zheng, Z., & Lu, G. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3443503 · Cited by 7
“In recent years, VO has attracted great interest from the computer vision and robotics communities [2], [3].”
BackgroundPerceptual Enhancement for Unsupervised Monocular Visual Odometry
Wang, Z., Shen, M., Liu, C., & Chen, Q. (2025) · International Journal of Control Automation and Systems · doi:10.1007/s12555-024-0495-y · Cited by 7
A data-driven based hybrid multi-branch framework for AUV navigation
Zhang, X., Sheng, L., He, B., & Lu, Y. P. (2025) · Ocean Engineering · doi:10.1016/j.oceaneng.2025.120675 · Cited by 8
Design, protection, perception and planning of mobile robots in nuclear power plants
De, Z., Jiang, Y. P., Liu, J., Xiong, G., & Zou, S. (2025) · Progress in Nuclear Energy · doi:10.1016/j.pnucene.2025.105821 · Cited by 7
Point-line feature-based vSLAM systems: A survey
Qu, H., Hu, Z., Zhao, Y., Lu, J., Ding, K., Liu, G., Chen, Y., & Shao, C. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127574 · Cited by 7
Zhou, F., Li, J., Dai, Y., Liu, L., Qin, H., Jiang, Y., Hong, S., Zhao, B. W., & Yang, T. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3230792 · Cited by 7
“SFM is also a popular method for training end-to-end models using deep-learning networks, such as FlowNet networks based on optical flow [25] and supervised network models for end-to-end 6-DOF pose estimation [26], [27].”
MethodPrinciples of Object Tracking and Mapping
Rambach, J. R., Pagani, A., & Stricker, D. (2023) · Springer handbooks · doi:10.1007/978-3-030-67822-7_3 · Cited by 7
A Biomorphic Whisker Sensor for Aerial Tactile Applications
Ye, C., de Croon, G., & Hamaza, S. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10610850 · Cited by 8
“First, directly regressing angles using L 1 = ∥ ∥ 1 or L 2 = ∥ ∥ 2 loss functions is a simple and effective approach [23].”
MethodIBD-SLAM: Learning Image-Based Depth Fusion for Generalizable SLAM
Yin, M., Wu, S., & Han, K. (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/cvpr52733.2024.01005 · Cited by 7
“Popular approaches include feature-based methods [16, 24], direct methods [11, 28], and, more recently, learning-based meth-ods [10, 44].”
MethodConstant Velocity 3D Convolution
Sekikawa, Y., Ishikawa, K., & Saitô, H. (2018) · IEEE Access · doi:10.1109/access.2018.2883340 · Cited by 7
“Learning spatiotemporal features from sequential data is crucial for applications such as action/activity recognition [15], visual odometry (VO) estimation [2], [18], [21], and policy learning [19].”
BackgroundUnsupervised Learning of Camera Pose with Compositional Re-estimation
Nabavi, S. S., Hosseinzadeh, M., Fahimi, R., & Wang, Y. (2020) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv45572.2020.9093495 · Cited by 7
“DeepVO [40] performs end-to-end visual odometry.”
BackgroundThe Optimized Distance Calculation Method with Stereo Camera for an Autonomous Tomato Harvesting
Buribayev, Z., Merembayev, T., Amirgaliyev, Y. N., & Miyachi, T. (2021) · IEEE International Conference on Smart Information Systems and Technologies (SIST) · doi:10.1109/sist50301.2021.9465944 · Cited by 7
Vehicle Trajectory Estimation Based on Fusion of Visual Motion Features and Deep Learning
Qu, L., & Dailey, M. N. (2021) · Sensors · doi:10.3390/s21237969 · Cited by 7
DBLD-SLAM: A Deep-Learning Visual SLAM System Based on Deep Binary Local Descriptor
Gu, X., Wang, Y., & Ma, T. (2021) · International Conference on Control, Automation and Information Sciences (ICCAIS) · doi:10.1109/iccais52680.2021.9624536 · Cited by 7
Exploiting Graph and Geodesic Distance Constraint for Deep Learning-Based Visual Odometry
Xu, F., Li, Q., Li, Q., Ding, K., & Zhu, J. (2022) · Remote Sensing · doi:10.3390/rs14081854 · Cited by 8
Attention Guided Unsupervised learning of Monocular Visual-inertial Odometry
Wang, Z., Zhu, Y., Lu, K., Freer, D., Wu, H., & Chen, H. (2022) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv51971.2022.9827359 · Cited by 7
An Unsupervised Approach for Simultaneous Visual Odometry and Single Image Depth Estimation
Lu, Y., & Lu, G. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9892294 · Cited by 7
“Wang et al. [27] trained a recurrent convolutional neural network on top of AlexNet to estimate the image depth and camera odometry.”
BackgroundTowards Specialized Hardware for Learning-based Visual Odometry on the Edge
Chen, S., & Mai, K. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9982046 · Cited by 8
A New Dense Hybrid Stereo Visual Odometry Approach
Liu, Z., Malis, E., & Martinet, P. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981814 · Cited by 7
“Recently, more and more end-to-end deep learning visual odometry approaches have been proposed, including supervised [5] and unsupervised [6] models.”
BackgroundEstimating Withdrawal Time in Colonoscopies
Katzir, L., Veikherman, D., Dashinsky, V., Goldenberg, R., Shimshoni, I., Rabani, N., Cohen, R., Kelner, O., Rivlin, E., & Freedman, D. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-25066-8_28 · Cited by 7
Moving Vehicle Detection with Real Time Speed Estimation Using Deep Learning Techniques
Pandilwar, V., & Kaur, N. (2023) · International Conference for Emerging Technology (INCET) · doi:10.1109/incet57972.2023.10170226 · Cited by 7
“DeepVO uses recurrent convolutional neural networks to learn the underlying motion patterns [14].”
BackgroundGPS Tracking Based on Stacked-Serial LSTM Network
Jin, X., Liu, S., Kong, J., Bai, Y., Su, T., & Ma, H. (2024) · Chinese journal of information fusion. · doi:10.62762/cjif.2024.361889 · Cited by 7
“In recent years, there has been significant progress in predicting the GPS position information of moving targets using neural networks [19–21].”
BackgroundTian, Y., & Compere, M. D. (2019) · IEEE International Conference on Artificial Intelligence and Virtual Reality (AIVR) · doi:10.1109/aivr46125.2019.00043 · Cited by 7
“This dataset has been widely accepted and its accuracy has been validated by many studies [10, 17, 18].”
MethodZhao, H., Bian, W., Yuan, B., & Tao, D. (2020) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2020/68 · Cited by 7
“DeepVO [Wang et al. , 2017] combines CNNs and Long Short-Term Memory Networks (LSTMs) to obtain pose estimation from image sequences by conducting sequential modelling.”
MethodLearning allocentric representations of space for navigation
Zhao, D., Si, B., & Li, X. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.10.013 · Cited by 7
Overview of Visual SLAM for Mobile Robots
Chen, W. R. (2021) · International Journal of Frontiers in Engineering Technology · doi:10.25236/ijfet.2021.030701 · Cited by 7
DynaNet: Neural Kalman Dynamical Model for Motion Estimation and PredictionPreprint
Chen, C., Lu, C. X., Wang, B., Trigoni, N., & Markham, A. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1908.03918 · Cited by 9
“As a consequence, in recent years, there has been an explosive growth in applying deep neural networks (DNNs) for motion estimation [9], [10], [11], [12], [13], [14], [15], [16], [17].”
BackgroundMethodSelf-citationSelf-Supervised Learning of Depth and Ego-motion with Differentiable Bundle AdjustmentPreprint
Shi, Y., Zhu, J. H., Fang, Y., Lien, K.-C., & Gu, J. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.13163 · Cited by 7
“[51] proposed a recurrent ConvNet architecture for learning monocular odometry from video sequences.”
BackgroundSelectFusion: A Generic Framework to Selectively Learn Multisensory Fusion.Preprint
Chen, C., Rosa, S., Lu, C. X., Trigoni, N., & Markham, A. (2019) · arXiv (Cornell University) · Cited by 7
Self-citationLeft/Right Invariant Lie Group Error for SINS/GNSS Tightly Coupled Vehicular Integrated Navigation
Wang, M., Cui, J., & Wu, W. (2025) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2025.3542441 · Cited by 6
Wei, Y., Lu, S., Lu, W., Xiong, R., & Wang, Y. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3547696 · Cited by 6
“Early works [5], [6] extract features from perspective images to regress poses but show limited performance.”
Self-Supervised Visual Odometry Based on Scene Appearance-Structure Incremental Fusion
Fu, F., Yang, J., Ma, J., & Zhang, J. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2025.3559077 · Cited by 6
“With the vigorous development of deep learning, the deep learning-based VO methods [7], [8], [9] leverage rich priors from learnable data to exhibit stronger robustness against the aforementioned challenges.”
DINO-VO: A Feature-Based Visual Odometry Leveraging a Visual Foundation Model
Azhari, M. B., & Shim, D. H. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3592065 · Cited by 9
“Restrictions apply.”
Seo, H.-I., Bae, J.-W., Kim, W.-Y., & Seo, D.-H. (2023) · IEEE Access · doi:10.1109/access.2023.3281970 · Cited by 6
“stereo, or monocular camera [3], [4], [5].”
Geometry-Aware Network for Unsupervised Learning of Monocular Camera’s Ego-Motion
Zhou, B., Xie, J., Jin, Z., & Kong, H. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3298715 · Cited by 5
“Supervised methods [6], [7], [29], [30], [31] have achieved robust performance but need expensive ground-truth labels as supervision.”
MethodSurvey of Deep Learning-Based Methods for FMCW Radar Odometry and Ego-Localization
Brune, M., Meisen, T., & Pomp, A. (2024) · Applied Sciences · doi:10.3390/app14062267 · Cited by 6
“This cutoff criterion was motivated by the famous paper by Wang et al. [83] using deep learning for visual odometry published in 2017.”
MethodA Global Pose and Relative Pose Fusion Network for Monocular Visual Odometry
Su, B., & Zang, T. (2024) · IEEE Access · doi:10.1109/access.2024.3439529 · Cited by 6
“Classic approaches such as DeepVO [1] use convolutional neural networks (CNNs) and recurrent neural networks (RNNs) for extracting and learning sequence features.”
MethodInfluentialXie, Y., Yang, J., Zhou, H., & Sun, F. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3477292 · Cited by 5
UL-VIO: Ultra-Lightweight Visual-Inertial Odometry with Noise Robust Test-Time Adaptation
Park, J., Chun, S. Y., & Seok, M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73036-8_24 · Cited by 6
“Estimating one’s ego-motion from camera images and inertial measurement unit (IMU) data sequences [11, 25], VIO is a crucial component in the autonomous navigation pipeline [33,37,40].”
Toward Specialized Learning-based Approaches for Visual Odometry: A Comprehensive Survey
Phan, T.-D., & Kim, G.-W. (2025) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-025-02245-0 · Cited by 9
“This architectural design allows the model to process sequential visual data more efficiently than traditional VO methods like ORB-SLAM2 [64] or sequence-based models such as DeepVO [166].”
Constant Velocity 3D Convolution
Sekikawa, Y., Ishikawa, K., Hara, K., Yoshida, Y., Suzuki, K., Sato, I., & Saitô, H. (2018) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2018.00047 · Cited by 6
“Learning spatiotemporal features from sequential data is crucial for applications such as action/activity recognition [12] and visual odometry (VO) estimation [14, 2, 15].”
BackgroundSemantic-only Visual Odometry based on dense class-level segmentation
Mahe, H., Marraud, D., & Comport, A. I. (2018) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr.2018.8545263 · Cited by 6
“[18] extend previous works encapsulating a CNN with stacked image pairs as input in a Recurrent Neural Network (RNN) that implicitly models sequential dynamics and relations.”
BackgroundNeural Network-Based Indoor Autonomously-Navigated AGV Motion Trajectory Data Fusion
Quan, Y., Huang, L., Ma, L., He, Y., & Wang, R. (2021) · Automatic Control and Computer Sciences · doi:10.3103/s0146411621040076 · Cited by 7
Visual Odometry Algorithm Based on Deep Learning
Liu, H., Huang, D. D., & Geng, Z. Y. (2021) · International Conference on Image, Vision and Computing (ICIVC) · doi:10.1109/icivc52351.2021.9526975 · Cited by 6
Vision-based speedometer regulates human walking
Takamuku, S., & Gomi, H. (2021) · iScience · doi:10.1016/j.isci.2021.103390 · Cited by 7
“Self-motion speed estimation frommonocular vision has only recently been explored in the field of computer vision (Li et al., 2018; Wang et al., 2017; Zhan et al., 2018) and further study on the primate visual odometry system may contribute to developing advanced algorithms for speed monitoring in…”
BackgroundEscudero, N., Costas, P., Hardt, M. W., & İnalhan, G. (2022) · Integrated Communications, Navigation and Surveillance Conference (ICNS) · doi:10.1109/icns54818.2022.9771519 · Cited by 6
Keypoint Heatmap Guided Self-Supervised Monocular Visual Odometry
Xiu, H., Liang, Y., & Zeng, H. (2022) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-022-01685-2 · Cited by 6
“With the development of deep learning technology and its successful application in the field of image recognition, many monocular visual odometry methods based on deep learning have been proposed [10, 12, 13, 26–29].”
MethodWang, C., Zhang, J., Zhao, Y. H., Zhou, Y., & Jiang, J. (2023) · Chinese Journal of Mechanical Engineering · doi:10.1186/s10033-023-00872-y · Cited by 6
Chen, J., Zhang, S., Li, Z., & Jin, X. (2023) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn54540.2023.10191248 · Cited by 6
“DeepVO [15] adopts a standard convolution-recurrent network for regression of pose, with the supervision of pose ground truth.”
BackgroundMethodIntrospective perception for mobile robots
Rabiee, S., & Biswas, J. (2023) · Artificial Intelligence · doi:10.1016/j.artint.2023.103999 · Cited by 6
Scale-Aware Monocular Visual Odometry and Extrinsic Calibration Using Vehicle Kinematics
Kim, C., Jang, Y., Kim, J., Kim, P., & Kim, H. J. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3309833 · Cited by 6
“Following the trend, a number of MVO attempts using deep learning are also introduced [25], [26], [27], [28], [29], [30].”
MethodQu, H., Zhang, L., Hu, X., He, X., Pan, X., & Chen, C. (2023) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2023.3342803 · Cited by 6
“Supervised learning based VO/VIO models require ground-truth pose labels to train the neural network [1], [2].”
BackgroundMethodMonocular Teach-and-Repeat Navigation using a Deep Steering Network with Scale Estimation
Zhao, C., Sun, L., Krajník, T., Duckett, T., & Yan, Z. (2021) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros51168.2021.9635912 · Cited by 6
“We can see that the performance of Scale-Transformer outperforms both supervised deep learning-based VO [4], [29] and self-supervised deep learning-based VO [10], [30].”
BackgroundMethodInfluentialAttention-based Long-term Modeling for Deep Visual Odometry
Xu, S., Xiong, H., Wu, Q., & Wang, Z. (2021) · International Conference on Digital Image Computing: Techniques and Applications (DICTA) · doi:10.1109/dicta52665.2021.9647140 · Cited by 6
Learning Multiplicative Interactions with Bayesian Neural Networks for Visual-Inertial OdometryPreprint
Shinde, K., Lee, J., Humt, M., Sezgin, A., & Triebel, R. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2007.07630 · Cited by 9
“Addition of LSTM after CNN similar to Wang et al. (2017) is made to the network in (Walch et al., 2017; Clark et al., 2017a) and the network is further improved by making the CNN part Bayesian to estimate uncertainty of localization in Kendall & Cipolla (2016).”
InfluentialDeep learning based monocular visual-inertial odometry
Esfahani, M. A. (2020) · doi:10.32657/10356/152280 · Cited by 6
Towards Interpretable Camera and LiDAR data fusion for Unmanned Autonomous Vehicles LocalisationPreprint
Tibebu, H., Silva, D., Artaud, C., Pina, R., & Shi, X. (2022) · Preprints.org · doi:10.20944/preprints202209.0276.v1 · Cited by 6
Learning-Based Heatmap-Guided Model for Monocular Visual Odometry
Lin, H.-C., Jhong, S.-Y., Lin, Y., Fang, J.-K., Chang, C., Lu, S.-Y., & Chen, Y.-Y. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3584151 · Cited by 5
“Traditional visual alignment calculation methods require careful design and tuning by adjusting the parameters to suit each application and its environmental conditions [6].”
InfluentialHybrid model navigation method for autonomous underwater vehicle
Zhang, X., He, B., Mu, P., & Zhang, D. (2022) · Ocean Engineering · doi:10.1016/j.oceaneng.2022.112027 · Cited by 5
Optimized Deep Learning for LiDAR and Visual Odometry Fusion in Autonomous Driving
Zhang, D., Peng, T., & Loomis, J. S. (2023) · IEEE Sensors Journal · doi:10.1109/jsen.2023.3324559 · Cited by 5
“To provide a comprehensive evaluation, we also test the deep learning-based visual odometry RCNN model [17] using the same dataset and compare it with our proposed system.”
MethodGRVINS: Tightly Coupled GNSS-Range-Visual-Inertial System
Lu, B.-X., Tsai, Y.-C., & Tseng, K.-S. (2024) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-023-02033-8 · Cited by 5
Hyperion – A Fast, Versatile Symbolic Gaussian Belief Propagation Framework for Continuous-Time SLAM
Hug, D., Alzugaray, I., & Chli, M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73404-5_13 · Cited by 5
Ghasemieh, A., & Kashef, R. F. (2024) · Sensors · doi:10.3390/s24248040 · Cited by 5
BEV-ODOM: Reducing Scale Drift in Monocular Visual Odometry with BEV Representation
Wei, Y., Lu, S., Han, F., Xiong, R., & Wang, Y. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10801996 · Cited by 5
“DeepVO [6] employs deep recurrent convolutional neural networks for end-to-end visual odometry, improving accuracy and robustness.”
Metadata
- Authors
- Wang, S., Clark, R., Wen, H., & Trigoni, N.
- Affiliations
- University of Oxford (GB)
- Venue
- 2017 IEEE International Conference on Robotics and Automation (ICRA)
- Publisher
- IEEE
- Published
- 2017-05
- Pages
- 2043-2050
- ISSN
- 1050-4729 (ISSN-L)
- OpenAlex
- W2598706937
- Cited by
- 970
- References
- 29
- Citations by year
- 2026: 63 · 2025: 120 · 2024: 104 · 2023: 123 · 2022: 130 · 2021: 131 · 2020: 125 · 2019: 116 · 2018: 52 · 2017: 6
- FWCI
- 143.15
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Proceedings - IEEE International Conference on Robotics and Automation/Proceedings (publishedVersion)
- arXiv (Cornell University) (publishedVersion)
- Warwick Research Archive Portal (University of Warwick) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Engineering and Physical Sciences Research Council (EP/M019918/1)
- 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
- monocular visual odometry, end-to-end learning, recurrent convolutional neural network, pose estimation, feature representation learning
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
