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
Flagged first, then by citations
Research works printed but not linked: 1
- zaremba, 2014, Learning to execute
Cited by
Sorted by citations
Items 601–700 of 969
Multimodal Fusion SLAM With Fourier Attention
Zhou, Y., Mei, G., Wang, Y., Wan, Y., & Poiesi, F. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3512252 · Cited by 4
“DeepVO [21] estimates camera poses from consecutive frames with an end-to-end framework, while DeepV2D [22] alternates between depth prediction and camera pose estimation, without relying on bundle adjustment.”
Implicit Self-Augmentation and Soft Dominance Prediction for Pedestrian Inertial Localization
Li, Y., Shi, Z., Hou, Y., Xie, L., Chen, H., Yan, Y., & Yin, E. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3554891 · Cited by 3
“D velocities v n from an input window w of IMU measurements in an end-to-end manner [10], [35], [36], [37], [38] where θ is the parameter set of latent neural networks; position p n can be obtained by the integration of velocity; p 0 n and v 0 n are initial states in the navigation coordinate; and…”
Dense Monocular SLAM in Real-Time With Structured Gaussian Representation
Liu, S., Wei, X., Zhao, C. K., Tian, A., & Du, B. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3583629 · Cited by 3
A Research on the Fusion of Semantic Segment Network and SLAM
Chen, Z., Zhang, W., Li, F. F., Shi, Y., Wang, Y., Nie, F., Zhu, C., & Huang, Q. (2019) · IEEE Workshop on Advanced Robotics and its Social Impacts (ARSO) · doi:10.1109/arso46408.2019.8948779 · Cited by 3
“[14], DeMon by Ummenhofer et al.”
Salient FlowNet and Decoupled LSTM Network for Robust Visual Odometry
Chen, M., Wang, C., & Liu, H. (2019) · IEEE International Conference on Robotics and Biomimetics (ROBIO) · doi:10.1109/robio49542.2019.8961700 · Cited by 3
RidgeSfM: Structure from Motion via Robust Pairwise Matching Under Depth Uncertainty
Graham, B., & Novotný, D. (2020) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv50981.2020.00075 · Cited by 3
“Initial approaches have focused solely on monocular depth estimation [10, 22, 46, 23, 24], or on estimating the camera pose [2, 20, 18, 19, 47].”
BackgroundSelf-supervised optical flow derotation network for rotation estimation of a spherical camera
Kim, D., Pathak, S., Moro, A., Yamashita, A., & Asama, H. (2020) · Advanced Robotics · doi:10.1080/01691864.2020.1857305 · Cited by 3
Point Cloud based Hierarchical Deep Odometry Estimation
Nowruzi, F. E., Kolhatkar, D., Kapoor, P., & Laganière, R. (2021) · International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS) · doi:10.5220/0010442900002932 · Cited by 3
Xu, Q., Bin, C., Li, X., Xixiang, L., & Tian, Y. (2021) · Saint Petersburg International Conference on Integrated Navigation Systems · doi:10.23919/icins43216.2021.9470863 · Cited by 3
Application and Analysis of Recurrent Convolutional Neural Network in Visual Odometry
Xu, D., & Zhang, Z. (2022) · IEEE International Conference on Mechatronics and Automation (ICMA) · doi:10.1109/icma54519.2022.9856336 · Cited by 3
“Because calculating the optical flow of RGB images is time-consuming and requires powerful hardware support, which also has a certain impact on the real-time performance of the system as a whole, Deep VO [10] and GCN [11] try to use Recurrent Convolutional Neural Networks to extract inter-frame information from image…”
MethodGreenPCO: An Unsupervised Lightweight Point Cloud Odometry Method
Kadam, P., Zhang, M., Gu, J., Liu, S., & Kuo, C. ‐. J. (2022) · IEEE International Workshop on Multimedia Signal Processing (MMSP) · doi:10.1109/mmsp55362.2022.9950022 · Cited by 3
“They are Twostream [16], DeepVO [21], PointNet [17], PointGrid [13] and DeepPCO [22].”
BackgroundMethodFault Resistant Odometry Estimation using Message Passing Neural Network
Dahal, P., Mentasti, S., Paparusso, L., Arrigoni, S., & Braghin, F. (2023) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv55152.2023.10186649 · Cited by 3
“With the recent development of the learning-based algorithms based on Deep Neural Networks (DNN), interest has shifted from odometry estimation using the 3D geometry and camera models to the learning-based pose estimators [1], [2], [3].”
MethodSong, K., & Chang, Y. C. (2018) · International Automatic Control Conference · doi:10.1109/cacs.2018.8606773 · Cited by 3
“This makes the system become complicated to setup[7].”
BackgroundEfficient and Robust Homography Estimation Using Compressed Convolutional Neural Network
Wang, G., You, Z., An, P., Yu, J., & Chen, Y. (2019) · Communications in computer and information science · doi:10.1007/978-981-13-8138-6_13 · Cited by 3
LightVO: Lightweight Inertial-Assisted Monocular Visual Odometry with Dense Neural Networks
Guo, Z., Yang, M., Chen, N., Xiao, Z., Yan, B., Lin, S., & Zhou, L. (2019) · Globecom · doi:10.1109/globecom38437.2019.9013757 · Cited by 3
“Wang et al. [19] propose a new end-to-end monocular VIO framework based on RCNN using deep recursive convolutional neural networks.”
MethodSelf-Supervised Learning of Visual Odometry
Song, L., & Luo, W. (2020) · International Conference on Information Science, Parallel and Distributed Systems (ISPDS) · doi:10.1109/ispds51347.2020.00009 · Cited by 3
“Optical flow contains rich image motion information, which is usually used as a self-supervised signal for VO model [2,25].”
BackgroundMethodVisual Landmark Learning Via Attention-Based Deep Neural Networks
You, M., Luo, C., & Zhou, H. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3103257 · Cited by 3
RGB-D Odometry for Autonomous Lawn Mowing
Ochman, M., Skoczeń, M., Krata, D., Panek, M., Spyra, K., & Pawłowski, A. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-87897-9_8 · Cited by 3
Body Temperature Prediction with Recurrent Neural Network and its Variants
Liang, Z., Wang, X., Chen, Z., & Luo, X. (2021) · International Conference on Intelligent Control and Information Processing · doi:10.1109/icicip53388.2021.9642166 · Cited by 3
Deep Weakly Supervised Positioning for Indoor Mobile Robots
Wang, R., Xu, X., Ding, L., Huang, Y., & Feng, C. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3138170 · Cited by 3
Szántó, M., Bogár, G. R., & Vajta, L. (2022) · Periodica Polytechnica Electrical Engineering and Computer Science · doi:10.3311/ppee.20437 · Cited by 3
“[32] showed that through training a VO network in the supervised way, it is easier to achieve real world metric results if the ground truth data is also measured in a real-world reference.”
BackgroundLu, G. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9892072 · Cited by 3
“The deep neural network is also applied in VO [24][35][38].”
MethodIQ-VIO: adaptive visual inertial odometry via interference quantization under dynamic environments
Zhang, H., Ye, F., Lai, Y., Li, K., & Xu, J. (2023) · Intelligent Service Robotics · doi:10.1007/s11370-023-00478-2 · Cited by 3
“In recent years, researchers demonstrate the feasibility of estimating pose directly using deep learning network, i.e., GeoNet [23] and DeepVO [24].”
BackgroundLine–Circle–Square (LCS): A multilayered geometric filter for edge-based detection
Tafrishi, S. A., Dai, X., & Kandjani, V. E. (2021) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2021.103732 · Cited by 3
D2VO: Monocular Deep Direct Visual Odometry
Jia, Q., Pu, Y., Chen, J., Cheng, J., Liao, C., & Yang, X. Y. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341313 · Cited by 3
“Some methods also proposed to estimate camera pose based on CNN. DeepVO [31] employed an RNN to predict camera pose end-to-end from input image sequence.”
MethodScene Motion Decomposition for Learnable Visual OdometryPreprint
Slinko, I., Vorontsova, A., Konokhov, F., Barinova, O., & Konushin, A. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1907.07227 · Cited by 3
“Excellent performance of deep learning-based methods for OF estimation [7, 13, 24] led to emergence of end-to-end VO algorithms utilizing OF [3, 5, 26].”
BackgroundMethodGraph-Based Thermal–Inertial SLAM With Probabilistic Neural NetworksPreprint
Saputra, M. R. U., Lu, C. X., de Gusmão, P. P. B., Wang, B., Markham, A., & Trigoni, N. (2021) · IEEE Transactions on robotics · doi:10.48550/arxiv.2104.07196 · Cited by 51
“2) Probabilistic Pose Regressor: Instead of directly predicting 6-DoF agent poses as in the previous work [12, 13], the probabilistic pose regressor estimates the parameters of MDN: mean (μ), variance (σ), and the mixing coefficients (α) as seen in Eq.”
BackgroundMethodInfluentialSelf-citationTwinvo: Unsupervised Learning of Monocular Visual Odometry Using Bi-Direction Twin Network
Cai, X., Zhang, L., Li, C., Li, G., & Li, T. H. (2020) · IEEE International Conference on Multimedia & Expo Workshops (ICMEW) · doi:10.1109/icmew46912.2020.9106039 · Cited by 2
“The DeepVO [2] firstly presents an end-to-end framework for monocular VO using deep Recurrent Convolutional Neural Networks (RCNNs).”
BackgroundMethodRelative pose estimation from panoramic images using a hybrid neural network architecture
Offermann, L. (2024) · Scientific Reports · doi:10.1038/s41598-024-75124-7 · Cited by 2
“Machine learning models that continuously estimate camera poses from image sequences are either trained supervised using pose data as ground truth 9,31,32 or unsupervised with video prediction as a surrogate task 33–35 .”
MethodQu, H., Wang, C., Xu, Y.-M., Zhang, L., Hu, X., & Chen, C. (2024) · Applied Sciences · doi:10.3390/app14219748 · Cited by 2
Ribeiro, L. M., Nedjah, N., & de Carvalho, P. V. R. (2025) · Applied Sciences · doi:10.3390/app15073433 · Cited by 2
“Recently, the state of the art in SLAM has incorporated deep-learning-based techniques, such as DeepVO [11] and CodeSLAM [12], which leverage neural networks to learn compact environment representations and improve the robustness of visual odometry.”
Song, J., Aouf, N., Rondao, D., Honvault, C., Mansilla, L. A., & Yang, Q. (2025) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2025.3580700 · Cited by 2
A dynamic SLAM algorithm based on improved YOLOv9S
Zhu, Q., Zhao, Y., Zhang, H., & Chen, W. (2025) · Applied Soft Computing · doi:10.1016/j.asoc.2025.113700 · Cited by 2
Tong, G., Zhang, Y., Li, Z., & Sun, J. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3606150 · Cited by 2
“Supervised models [6] require a large amount of accurately labeled ground truth data for pose and depth as supervision signals.”
FPEVO: Fused point-edge visual odometry for low-structured and low-textured scenes
Brown, D., Grobler, H., & de Villiers, J. P. (2025) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2025.104599 · Cited by 2
SwinFVO: Self-Supervised Visual Odometry With Enhanced Global Spatiotemporal Perception
Song, R., Li, R., Xiao, Z., & Yan, B. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3637304 · Cited by 2
“The emergence of deep learning (DL) methodologies in computer vision (CV) has provided e ff ective solutions for VO tasks [8], [9].”
A Robust and Efficient Visual-Inertial SLAM Using Hybrid Point-Line Features
Wen, S., Tan, S., Liu, X., Li, M., & Liu, H. H. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3648610 · Cited by 2
Learning-Based Multi-Robot Active SLAM: A Conceptual Framework and Survey
Lv, B., & Duan, S. (2026) · Applied Sciences · doi:10.3390/app16031412 · Cited by 3
A Novel Visual Odometry Network Based on Short Sequence Images With Local Constraint Loss Function
Li, S., & He, Z. (2026) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2026.3660434 · Cited by 2
“The proposed loss function does not, like those in [3], [32], [30], [23], and [36], treat the translation and rotation predictions as two independent parts and ignore the relationship between them.”
InfluentialM 2 EIT: Multi-Domain Mixture of Experts for Robust Neural Inertial Tracking
Li, Y., Xu, Y., Chen, C., Shi, Z., Chen, W., Xie, L., Chen, H., & Yin, E. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02619 · Cited by 2
“The goal of the NIT G is to regress velocities v n from X in an end-to-end manner [11, 25, 28, 33]. where θ denotes the parameters of the latent neural networks; p n 0 and v n 0 are initial states.”
Dynamic Object-Aware Monocular Visual Odometry With Local And Global Information Aggregation
Wan, Y., Gao, W., Han, S., & Wu, Y. (2020) · Proceedings - International Conference on Image Processing · doi:10.1109/icip40778.2020.9190930 · Cited by 2
“In DeepVO[7], features extracted from image sequences are directly fed into LSTM for temporal modelling.”
BackgroundMethodInfluentialENCODE: a dEep poiNt Cloud ODometry nEtwork
Zhang, Y., Wang, L., Fu, C., Dai, Y., & Dolan, J. M. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9562024 · Cited by 2
“In the visual odometry domain, deep learning-based methods have achieved promising results, including monocular visual odometry [21], stereo visual odometry [22] and unsupervised visual odometry [23], [24].”
MethodStereo Visual Odometry with Information Enhancement at Feature Points
Liu, S., & Ma, B. (2021) · International Conference on Big Data & Artificial Intelligence & Software Engineering (ICBASE) · doi:10.1109/icbase53849.2021.00129 · Cited by 2
“Wang et al. [40] proposed a recurrent network for learning VO from image sequences.”
BackgroundRobust Visual-Inertial Odometry Based on Deep Learning and Extended Kalman Filter
Zuo, S., Shen, K., & Zuo, J. (2021) · China Automation Congress (CAC) · doi:10.1109/cac53003.2021.9727349 · Cited by 2
Assessment of CNN-Based Models for Odometry Estimation Methods with LiDAR
Clavijo, M., Jiménez, F., Serradilla, F., & Díaz-Álvarez, A. (2022) · Mathematics · doi:10.3390/math10183234 · Cited by 2
VAE-TCN hybrid model for KPI Anomaly Detection
Wu, B., Xu, Q., Yao, Z., Tu, Y., & Chen, Y. (2022) · Asia-Pacific Network Operations and Management Symposium (APNOMS) · doi:10.23919/apnoms56106.2022.9919985 · Cited by 3
Unsupervised Monocular Visual Odometry with Lightweight Depth Architecture
Pan, L., & Liu, Z. (2023) · Lecture notes in electrical engineering · doi:10.1007/978-981-19-6613-2_228 · Cited by 2
Kou, K. C., Yang, G., Zhang, W., Wang, C., Yao, Y., & Zhou, X. (2023) · Communications in computer and information science · doi:10.1007/978-981-99-0301-6_4 · Cited by 2
DTR-Map: A Digital Twin-Enabled Real-Time Mapping System Based on Multi-View Stereo
Fu, W., Lv, B., & Tong, S. (2023) · China Automation Congress (CAC) · doi:10.1109/cac59555.2023.10450582 · Cited by 2
“DeepVO [10] utilizes CNNs for image feature extraction and models temporal sequences with RNNs. GeoNet [6] joint estimation of optical flow, depth, and Camera Motion.”
MethodVisual SLAM Algorithm Based on Deep Learning Feature Extraction and Mismatch Elimination
Wu, Z., Fan, Y., & Wang, G. (2023) · Intelligent Autonomous Systems · doi:10.1109/icoias61634.2023.00011 · Cited by 2
“DeepVO [9] extracts features through convolutional neural network ( CNN ) and learns continuous pose estimation on LSTM. GEN-SLAM [10] proposes a system based on deep learning, which can realize pose and depth estimation between image frames through monocular and depth cameras.”
BackgroundVIOFormer: Advancing Monocular Visual-Inertial Odometry Through Transformer-Based Fusion
Zhao, J., Zhu, Y., Huang, Y., Qiao, X., Guo, M., He, H., & Li, Y. (2024) · Learning and analytics in intelligent systems · doi:10.1007/978-3-031-56521-2_2 · Cited by 2
Hybrid self-supervised monocular visual odometry system based on spatio-temporal features
Yuan, S., Zhang, J., Lin, Y., & Yang, L. (2024) · Electronic Research Archive · doi:10.3934/era.2024163 · Cited by 2
“The Towards end-to-end visual odometry with deep recurrent convolutional neural networks (DeepVO) proposed by Wang et al. [17] implemented an end-to-end visual odometry framework.”
Testing the real-time performance of a monocular visual odometry method for a wheeled robot
Saleem, H., & Malekian, R. (2024) · IEEE International Systems Conference (SysCon) · doi:10.1109/syscon61195.2024.10553447 · Cited by 2
“TartanVO demonstrates superior performance on the Tar-tanAir dataset and comparable results on the KITTI and Eu-RoC datasets when compared to other state-of-the-art methods, such as ORB-SLAM [22], DeepVO [23], and UnDeepVO [24].”
ResultEnhancing Visual Odometry with Estimated Scene Depth: Leveraging RGB-D Data with Deep Learning
Kostusiak, A., & Skrzypczyński, P. (2024) · Electronics · doi:10.3390/electronics13142755 · Cited by 3
Learning Visual-Inertial Odometry With Robocentric Iterated Extended Kalman Filter
Nguyen, K., Tran, D. T., Pham, V., Nguyen, D. T., Inoue, K., Lee, J., & Nguyen, A. Q. (2024) · IEEE Access · doi:10.1109/access.2024.3440182 · Cited by 2
“Toaddress these limitations, researchers have proposed employing deep learning methods to enhance the accuracy and efficiency of ego-motion estimation [5], [6], [7].”
BackgroundMethodAchieving High-Precision Localization in Self-driving Cars Using Real-Time Visual-Based Systems
Viet, P. T., & Hung, P. D. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-71315-6_32 · Cited by 2
Weakly Supervised End2End Deep Visual Odometry
Abouee, A., Ravi, A., Hinneburg, L., Dziwulski, M., Ölsner, F., Hess, J., Milz, S., & Mäder, P. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00091 · Cited by 2
“Supervised Methods : With the development of deep neural networks, approaches based on end-to-end learning have been proposed to solve the VO problem [22, 26, 27, 30, 31].”
BackgroundLiu, H., Gong, Z., Shen, J., Li, Y., & Long, Q. (2024) · Micromachines · doi:10.3390/mi15111362 · Cited by 2
Visual and Inertial Sensor Fusion Approach for Visual Inertial Odometry Used in Vehicle Localization
Jo, H., & Yoon, S. W. (2024) · Proceedings of IEEE Sensors/Proceedings of IEEE Sensors ... · doi:10.1109/sensors60989.2024.10785070 · Cited by 2
Semi-supervised pipe video temporal defect interval localization
Huang, Z., Pan, G., Kang, C., & Lv, Y. (2025) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.13403 · Cited by 2
“However, due to the unavailability of intrinsic parameters for the cameras used to capture the dataset’s videos, this study employs the pretrained visual odometry network DeepVO (S. Wang et al. 2017), which does not require camera intrinsics.”
OFVO: A Visual Odometry Designed for Motion Trajectory Estimation of Autonomous Vehicles
Wei, F., & Wang, W. (2024) · International Conference on Robotics, Automation and Artificial Intelligence (RAAI) · doi:10.1109/raai64504.2024.10949531 · Cited by 2
“Representative solutions include DeepVO [14], GeoNet [15], CNN-SLAM [16], etc.”
GD-MVO: Monocular Geometric Visual Odometry Using Image Depth Prediction
Li, Z., Chen, J., Waheed, A., Wang, F., & Fan, Y. (2025) · IEEE Access · doi:10.1109/access.2025.3562956 · Cited by 2
“Early research focused on supervised learning, utilizing ground truth data for tasks such as relative attitude estimation [8], [9] and depth prediction [10].”
Extreme Two-View Geometry From Object Poses with Diffusion Models
Sun, Y., Sun, C., Liu, Y., Ma, Y., & Yiu, S. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-5812-1_16 · Cited by 2
3D measurement of dynamic structures using monocular camera
Kobayashi, K., & Fuse, T. (2025) · Automation in Construction · doi:10.1016/j.autcon.2025.106240 · Cited by 2
Lu, G. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.01539 · Cited by 3
“As supervised methods for structure-from-motion, Wang et al. [71] proposed a recurrent convolutional neural network to learn monocular SfM from videos.”
A Robust and Accurate Stereo SLAM Based on Learned Feature Extraction and Matching
Liu, Y., Jia, T., Guo, S., Liu, Y., Zhang, K., & Du, Z. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3614846 · Cited by 2
“The state-of-the-art technology in this category is represented by [15] and [16].”
Lee, D., Nam, H., Jang, I., & Shim, D. H. (2025) · ETRI Journal · doi:10.4218/etrij.2025-0011 · Cited by 2
“Recently, CNN-based odometry methods have been proposed, and these learning-based studies have originated from research in visual odometry [38, 39].”
Exploitation of road signalling for localization refinement of autonomous vehicles
D’Orazio, L., Conci, N., & Stoffella, F. (2018) · AEIT International Conference of Electrical and Electronic Technologies for Automotive (AEIT AUTOMOTIVE) · doi:10.23919/eeta.2018.8493200 · Cited by 2
“DeepVO is a framework based on Recurrent Neural Networks (RNNs), for end-to-end visual odometry relying on a single monocular camera [23].”
MethodLocal Pose optimization with an Attention-based Neural Network
Liu, Y., Wang, H., Xu, F., Wang, Y., Chen, W., & Tang, Q. (2019) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros40897.2019.8968524 · Cited by 2
“Deepvo [18] estimates ego-motion by utilizing a recurrent convolutional network which can transmit long term information.”
BackgroundRigid-aware self-supervised GAN for camera ego-motion estimation
Lin, L., Luo, W., Yan, Z., & Zhou, W. (2022) · Digital Signal Processing · doi:10.1016/j.dsp.2022.103471 · Cited by 2
Descriptor-Length Reduction Using Low-Variance Filter for Visual Odometry
Kalambe, S. S., Rufus, E., Karar, V., & Poddar, S. (2019) · Advances in intelligent systems and computing · doi:10.1007/978-981-32-9291-8_1 · Cited by 2
Obstacle Avoidance Using Stereo Vision and Deep Reinforcement Learning in an Animal-like Robot
Ling, F., Jiménez-Rodríguez, A., & Prescott, T. J. (2019) · IEEE International Conference on Robotics and Biomimetics (ROBIO) · doi:10.1109/robio49542.2019.8961639 · Cited by 2
“Deep learning is playing an increasingly importance role in robotics and computer vision where it is being used to support capabilities such as planning path for collision avoidance and navigation robotics [7,8,9,16], however, standard supervised learning approaches typically require a large amount of manually labelle…”
BackgroundReliable Localization Using Multi-sensor Fusion for Automated Valet Parking Applications
Sakr, M. M., Moussa, A., Abdelfatah, W. F., Elsheikh, M., & El‐Sheimy, N. (2020) · Lecture notes in electrical engineering · doi:10.1007/978-981-15-3707-3_68 · Cited by 2
da Silva, A. M. R., Angonese, A. T., & Rosa, P. F. F. (2020) · Latin American Robotics Symposium, Brazilian Symposium on Robotics and Workshop on Robotics in Education (LARS/SBR/WRE) · doi:10.1109/lars/sbr/wre51543.2020.9306931 · Cited by 2
“Techniques like Convolutional Neural Network (CNN) [4] and Recurrent Neural Network (RNN) [5] provide excellent results in the Feature Extraction process and classification patterns in visual data and in VSLAM tasks [6], [7].”
MethodInformation-Theoretic Odometry Learning
Zhang, S., Zhang, J., & Tao, D. (2022) · International Journal of Computer Vision · doi:10.1007/s11263-022-01659-9 · Cited by 2
Incorporating learnt local and global embeddings into monocular visual SLAM
Huang, H., Ye, H., Sun, Y., Wang, L., & Liu, M. (2021) · Autonomous Robots · doi:10.1007/s10514-021-10007-8 · Cited by 2
“While some existing works generally recover pose in an end-to-end manner [16,17], limiting the generalization ability for different scenarios, we consider a more practical approach to utilize the deep feature.”
Monocular weakly supervised depth and pose estimation method based on multi-information fusion
Zhang, Z., Qiao, J., & Lin, S. (2021) · Kuwait Journal of Science · doi:10.48129/kjs.12929 · Cited by 1
Visual inertial system for ROV positioning
Castro, F., Marques, F. M., Costa, P. M. V. D., & Parente, M. A. (2020) · Technical Papers ... Rio Oil & Gas · doi:10.48072/2525-7579.rog.2020.403 · Cited by 2
A Comparison of Deep Learning-Based Monocular Visual Odometry Algorithms
Jeong, E., Lee, J., & Kim, P. (2022) · Lecture notes in electrical engineering · doi:10.1007/978-981-19-2635-8_68 · Cited by 2
Learning-Based Gaussian Belief Propagation for Bundle Adjustment in Visual SLAM
Feng, Y.-S., Chen, J., Wang, H.-C., Huang, C., & Chern, J. (2022) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps56602.2022.10008583 · Cited by 2
“The proposed method is trained under an end-to-end learning mechanism similar to DeepVO [18] with grid search [19] to determine the hyperparameters.”
MethodBenchmarking Pedestrian Odometry: The Brown Pedestrian Odometry Dataset (BPOD)
Charatan, D., Fan, H., & Kimia, B. B. (2022) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv57658.2022.00076 · Cited by 2
“egories, namely, feature-based methods [24], direct methods [6], [7] and deep learning methods [25], [26], [27], [28], and certain datasets can be more suitable for one or the other.”
MethodDeep 4D Automotive Radar-Camera Fusion Odometry with Cross-Modal Transformer Fusion
Lu, S., Zhuo, G., Xiong, L., Zhou, M., & Lu, X. (2023) · SAE International Journal of Advances and Current Practices in Mobility · doi:10.4271/2023-01-7040 · Cited by 2
Visual Simultaneous Localisation and Mapping Methodologies
Bouhamatou, Z., & Abdessemed, F. (2024) · Acta Mechanica et Automatica · doi:10.2478/ama-2024-0049 · Cited by 2
Perception and Decision-Making of Autonomous Systems in the Era of Learning: An Overview.
Tang, Y. S., Zhao, C., Wang, J., Zhang, C., Sun, Q., & Qian, F. (2020) · Cited by 2
Virtual testbed for monocular visual navigation of small unmanned aircraft systems
Kim, K. C., Leishman, R. C., & Nykl, S. L. (2020) · The Journal of Defense Modeling and Simulation Applications Methodology Technology · doi:10.1177/1548512920954545 · Cited by 2
Greyvensteyn, I. (2020) · Queensland University of Technology · doi:10.5204/thesis.eprints.204294 · Cited by 2
Experimental Comparison of Visual and Single-Receiver GPS OdometryPreprint
Congram, B., & Barfoot, T. D. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.02122 · Cited by 2
“Recently, deep-learning approaches [18], [19] to VO have gained interest, though feature-based methods still remain relevant.”
Enhancing Multi-Robot Perception via Learned Data AssociationPreprint
Glaser, N., Liu, Y., Tian, J., & Kira, Z. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.00769 · Cited by 2
“These assumptions are typically baked into the architectures of optical flow models [7, 20, 17] and visual odometry models [4, 18].”
A Survey of Computer Vision Algorithm Applied in Driverless Robotic Vehicles with Sensing CapabilityPreprint
Chen, L., Li, G., Xie, W., Tan, J., Li, Y., Pu, J., Chen, L., Gan, D., & Shi, W. (2024) · Preprints.org · doi:10.20944/preprints202405.1051.v1 · Cited by 2
Unsupervised monocular visual odometry via combining instance and RGB information
Yue, M., Fu, G., Wu, M., Gu, H., & Yao, E. (2022) · Applied Optics · doi:10.1364/ao.452378 · Cited by 1
A Compensation Method for Lost Trajectories of ORB-SLAM3 with DROID-VO
Liu, K.-R., Fan, Y., Su, J., Wu, L., & Zhang, K.-X. (2023) · 2023 3rd International Conference on Computer, Control and Robotics (ICCCR) · doi:10.1109/icccr56747.2023.10194078 · Cited by 1
“The most representative of supervised VO was DEEPVO[18], which extracted visual features from pairs of images via a convolutional neural network and passed these features through recurrent neural networks to model temporal correlation.”
MethodSequentially trained, shallow neural networks for real-time 3D odometry
Rodriguez, F., Muminov, B., & Vuong, L. T. (2023) · SPIE Conference on Artificial Intelligence for Security and Defence Applications · doi:10.1117/12.3005250 · Cited by 1
“17–21 Supervised 22 and unsupervised 23 learning approaches have been demonstrated to predict the speed, depth, and position of system objects, even with a monocular imaging system.”
BackgroundMelikyan, V. S., Grigoryan, M. T., Avetisyan, A., & Khachatryan, T. (2023) · IEEE East-West Design & Test Symposium (EWDTS) · doi:10.1109/ewdts59469.2023.10297049 · Cited by 1
“For instance, in [3], a novel end-to-end framework for VO based on deep learning is proposed.”
BackgroundAI-Based Visual Odometry Implementation on an Embedded System
Büyüksolak, O., & Güneş, E. O. (2023) · National Conference on Electrical and Electronics Engineering (ELECO) · doi:10.1109/iceee59925.2023.00016 · Cited by 1
“One of the early examples of this approach was introduced by the Wang et al. [6], named as DeepVO.”
MethodHNN-Transformer Integrated Network for Estimating Robot Position
Kang, C. H., & Kim, S. Y. (2023) · International Conference on Control, Automation and Systems (ICCAS) · doi:10.23919/iccas59377.2023.10316909 · Cited by 1
“For example, convolutional neural networks (CNNs) have been used for image-based localization tasks, while recurrent neural networks (RNNs) have proven effective at processing sequential data from odometry or inertial measurement units [4].”
MethodDai, M., Lai, X., Zhang, C., Pan, X., Ding, H., Zhang, S., Li, H., & Dai, Y. (2023) · Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC) · doi:10.1109/ipec57296.2023.00022 · Cited by 1
Wang, Z., Shen, M., & Chen, Q. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-18382-x · Cited by 1
Fast Visual-Inertial Odometry with Adaptive Feature Coupling
Ma, Z., Xiao, J., Wang, F., & Jiang, P. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-0792-1_14 · Cited by 1
Ajewole, O. K., Anjuwon, O., Ezugwu, C. A. K., Adetiba, G., Adewuyi, I. A., & Abiola, J. O. (2024) · IEEE Nigeria International Conference on Disruptive Technologies for Sustainable Development (NIGERCON) · doi:10.1109/nigercon62786.2024.10927138 · Cited by 1
Self-Supervised Learning of Camera Ego-Motion From Optical Flow
Xie, C., Hao, N., Yao, H., & He, F. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3577830 · Cited by 1
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
