Home

Awesome

Deep Learning for Localization and Mapping

image This repository is a collection of deep learning based localization and mapping approaches. A survey on Deep Learning for Visual Localization and Mapping is offered in the following paper:

Deep Learning for Visual Localization and Mapping: A Survey

Changhao Chen, Bing Wang, Chris Xiaoxuan Lu, Niki Trigoni and Andrew Markham

IEEE Transactions on Neural Networks and Learning Systems [PDF]

A survey on Deep Learning for Inertial Positioning is offered in the following paper:

Deep Learning for Inertial Positioning: A Survey

Changhao Chen, Xianfei Pan

IEEE Transactions on Intelligent Transportation Systems [PDF]

Previous Version.

A Survey on Deep Learning for Localization and Mapping: Towards the Age of Spatial Machine Intelligence

Changhao Chen, Bing Wang, Chris Xiaoxuan Lu, Niki Trigoni and Andrew Markham

arXiv:2006.12567 [PDF]

News

Update: Jun-22-2020

Update: Aug-30-2023

Update: Mar-13-2024

TO DO

Category

If you find this repository useful, please cite our paper:

@misc{chen2020survey,
title={A Survey on Deep Learning for Localization and Mapping: Towards the Age of Spatial Machine Intelligence},
author={Changhao Chen and Bing Wang and Chris Xiaoxuan Lu and Niki Trigoni and Andrew Markham},
year={2020},
eprint={2006.12567},
archivePrefix={arXiv},
primaryClass={cs.CV}
}

Categorized by Topic

*The Date in the table denotes the publication date (e.g. date of conference).

Odometry Estimation

Visual Odometry

ModelsDatePublicationPaperCode
Konda et al.2015VISAPPLearning visual odometry with a convolutional network
Costante et al.2016RA-LExploring Representation Learning With CNNs for Frame-to-Frame Ego-Motion Estimation
Backprop KF2016NeurIPSBackprop KF: Learning Discriminative Deterministic State Estimators
DeepVO2017ICRADeepVO: Towards End-to-End Visual Odometry with Deep Recurrent Convolutional Neural Networks
SfmLearner2017CVPRunsupervised learning of depth and ego-motion from videoTF PT
Yin et al.2017ICCVScale Recovery for Monocular Visual Odometry Using Depth Estimated With Deep Convolutional Neural Fields
UnDeepVO2018ICRAUnDeepVO: Monocular Visual Odometry through Unsupervised Deep Learning
Barnes et al.2018ICRADriven to Distraction: Self-Supervised Distractor Learning for Robust Monocular Visual Odometry in Urban Environments
GeoNet2018CVPRGeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera PoseTF
Zhan et al.2018CVPRUnsupervised Learning of Monocular Depth Estimation and Visual Odometry with Deep Feature ReconstructionCaffe
DPF2018RSSDifferentiable Particle Filters: End-to-End Learning with Algorithmic PriorsTF
Yang et al.2018ECCVDeep Virtual Stereo Odometry: Leveraging Deep Depth Prediction for Monocular Direct Sparse Odometry
Zhao et al.2018IROSLearning monocular visual odometry with dense 3d mapping from dense 3d flow
Turan et al.2018IROSUnsupervised Odometry and Depth Learning for Endoscopic Capsule Robots
Struct2Depth2019AAAIDepth Prediction Without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular VideosTF
Saputra et al.2019ICRALearning monocular visual odometry through geometry-aware curriculum learning
GANVO2019ICRAGANVO: Unsupervised deep monocular visual odometry and depth estimation with generative adversarial networks
CNN-SVO2019ICRACNN-SVO: Improving the Mapping in Semi-Direct Visual Odometry Using Single-Image Depth PredictionROS
Li et al.2019ICRAPose graph optimization for unsupervised monocular visual odometry
Xue et al.2019CVPRBeyond tracking: Selecting memory and refining poses for deep visual odometry
Wang et al.2019CVPRRecurrent neural network for (un-) supervised learning of monocular video visual odometry and depth
Li et al.2019ICCVSequential adversarial learning for self-supervised deep visual odometry
Saputra et al.2019ICCVDistilling knowledge from a deep pose regressor network
Gordon et al.2019ICCVDepth from videos in the wild: Unsupervised monocular depth learning from unknown camerasTF
Koumis et al.2019IROSEstimating Metric Scale Visual Odometry from Videos using 3D Convolutional Networks
Bian et al.2019NeurIPSUnsupervised Scale-consistent Depth and Ego-motion Learning from Monocular VideoPT
D3VO2020CVPRD3VO: Deep Depth, Deep Pose and Deep Uncertainty for Monocular Visual Odometry
Jiang et al.2020CVPRJoint Unsupervised Learning of Optical Flow and Egomotion with Bi-Level Optimization

Visual-Inertial Odometry

ModelsDatePublicationPaperCode
VINet2017AAAIVINet: Visual-Inertial Odometry as a Sequence-to-Sequence Learning Problem
VIOLearner2019TPAMIUnsupervised deep visual-inertial odometry with online error correction for rgb-d imagery
SelectFusion2019CVPRSelective Sensor Fusion for Neural Visual-Inertial Odometry
DeepVIO2019IROSDeepVIO: Self-supervised deep learning of monocular visual inertial odometry using 3d geometric constraints

Inertial Odometry

ModelsDatePublicationPaperCode
IONet2018AAAIIONet: Learning to Cure the Curse of Drift in Inertial Odometry
RIDI2018ECCVRIDI: Robust IMU Double IntegrationPy
Wagstaff et al.2018IPINLSTM-Based Zero-Velocity Detection for Robust Inertial NavigationPT
Cortes et al.2019MLSPDeep Learning Based Speed Estimation for Constraining Strapdown Inertial Navigation on Smartphones
MotionTransformer2019AAAIMotionTransformer: Transferring Neural Inertial Tracking between Domains
AbolDeepIO2019TITSAbolDeepIO: A Novel Deep Inertial Odometry Network for Autonomous Vehicles
Brossard et al.2019ICRALearning wheel odometry and imu errors for localization
OriNet2019RA-LOriNet: Robust 3-D Orientation Estimation With a Single Particular IMUPT
L-IONet2020IoT-JDeep Learning based Pedestrian Inertial Navigation: Methods, Dataset and On-Device Inference

LIDAR Odometry

ModelsDatePublicationPaperCode
Velas et al.2018ICARSCCNN for IMU Assisted Odometry Estimation using Velodyne LiDAR
LO-Net2019CVPRLO-Net: Deep Real-time Lidar Odometry
DeepPCO2019IROSDeepPCO: End-to-End Point Cloud Odometry through Deep Parallel Neural Network
Valente et al.2019IROSDeep sensor fusion for real-time odometry estimation

Mapping

Geometric Mapping

Depth Representation
ModelsDatePublicationPaperCode
Eigen et al.2014NeurIPSDepth Map Prediction from a Single Image using a Multi-Scale Deep Network
Liu et al.2015TPAMILearning depth from single monocular images using deep convolutional neural fields
Garg et al.2016ECCVUnsupervised cnn for single view depth estimation: Geometry to the rescue
Demon2017CVPRDemon: Depth and motion network for learning monocular stereo
Godard et al.2017CVPRUnsupervised monocular depth estimation with left-right consistency
Wang et al.2018CVPRLearning depth from monocular videos using direct methods
Voxel Representation
ModelsDatePublicationPaperCode
SurfaceNet2017CVPRSurfaceNet: An End-to-end 3D Neural Network for Multiview Stereopsis
Dai et al.2017CVPRShape completion using 3d-encoder-predictor cnns and shape synthesis
Hane et al.20173DVHierarchical surface prediction for 3d object reconstruction
OctNetFusion20173DVOctnetfusion: Learning depth fusion from data
OGN2017ICCVOctree generating networks: Efficient convolutional architectures for high-resolution 3d outputs
Kar et al.2017NeurIPSLearning a multi-view stereo machine
RayNet2018CVPRRayNet: Learning Volumetric 3D Reconstruction with Ray Potentials
Point Representation
ModelsDatePublicationPaperCode
Fan et al.2017CVPRA point set generation network for 3d object reconstruction from a single image
Mesh Representation
ModelsDatePublicationPaperCode
Ladicky et al.2017ICCVFrom point clouds to mesh using regression
Mukasa et al.2017ICCVW3d scene mesh from cnn depth predictions and sparse monocular slam
Wang et al.2018ECCVPixel2mesh: Generating 3d mesh models from single rgb images
Groueix et al.2018CVPRAtlasNet: A Papier-Mâché Approach to Learning 3D Surface Generation
Scan2Mesh2019CVPRScan2mesh: From unstructured range scans to 3d meshes
Bloesch et al.2019ICCVLearning meshes for dense visual SLAM

Semantic Mapping

ModelsDatePublicationPaperCode
SemanticFusion2017ICRASemanticfusion: Dense 3d semantic mapping with convolutional neural networks
DA-RNN2017RSSDA-RNN: Semantic mapping with data associated recurrent neural networks
Ma et al.2017IROSMulti-view deep learning for consistent semantic mapping with rgb-d cameras
Sunderhauf et al.2017IROSMeaningful maps with object-oriented semantic mapping
Fusion++20183DVFusion++: Volumetric object-level SLAM
Grinvald et al.2019RA-LVolumetric instance-aware semantic mapping and 3d object discovery
PanopticFusion2019IROSPanopticfusion: Online volumetric semantic mapping at the level of stuff and things

General Mapping

ModelsDatePublicationPaperCode
Mirowski et al.2017ICLRLearning to navigate in complex environments
Zhu et al.2017ICRATarget-driven visual navigation in indoor scenes using deep reinforcement learning
Eslami et al.2018ScienceNeural scene representation and rendering
CodeSLAM2018CVPRCodeSLAM — Learning a Compact, Optimisable Representation for Dense Visual SLAM
Mirowski et al.2018NeurIPSLearning to navigate in cities without a map
SRN2019NeurIPSScene representation networks: Continuous 3d-structure-aware neural scene representations
Tobin et al.2019NeurIPSGeometry-aware neural rendering
Lim et al.2019NeurIPSNeural multisensory scene inference

Global Localization

2D-to-2D Localization

Implicit Map Based Localization
ModelsDatePublicationPaperCode
PoseNet2015ICCVPoseNet: A Convolutional Network for Real-Time 6-DOF Camera Relocalization
Bayesian PoseNet2016ICRAModelling uncertainty in deep learning for camera relocalization
BranchNet2017ICRADelving deeper into convolutional neural networks for camera relocalization
VidLoc2017CVPRVidLoc: A Deep Spatio-Temporal Model for 6-DoF Video-Clip Relocalization
Geometric PoseNet2017CVPRGeometric loss functions for camera pose regression with deep learning
Naseer et al.2017IROSDeep Regression for Monocular Camera-based 6-DoF Global Localization in Outdoor Environments
LSTM-PoseNet2017ICCVImage-based localization using lstms for structured feature correlation
Hourglass PoseNet2017ICCV WorkshopsImage-based localization using hourglass networks
VLocNet2018ICRADeep auxiliary learning for visual localization and odometry
MapNet2018CVPRGeometry-Aware Learning of Maps for Camera Localization
SPP-Net2018BMVCSynthetic view generation for absolute pose regression and image synthesis
GPoseNet2018BMVCA hybrid probabilistic model for camera relocalization
VLocNet++2018RA-LVlocnet++: Deep multitask learning for semantic visual localization and odometry
Xue et al.2019ICCVLocal supports global: Deep camera relocalization with sequence enhancement
Huang et al.2019ICCVPrior guided dropout for robust visual localization in dynamic environments
Bui et al.2019ICCVWAdversarial networks for camera pose regression and refinement
GN-Net2020RA-LGN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization
AtLoc2020AAAIAtLoc: Attention Guided Camera Localization
Explicit Map Based Localization
ModelsDatePublicationPaperCode
Laskar et al.2017ICCV WorkshopsCamera Relocalization by Computing Pairwise Relative Poses Using Convolutional Neural Network
DELS-3D2018CVPRDels-3d: Deep localization and segmentation with a 3d semantic map
AnchorNet2018BMVCImproved visual relocalization by discovering anchor points
RelocNet2018ECCVRelocNet: Continuous Metric Learning Relocalisation using Neural Nets
CamNet2019ICCVCamnet: Coarse-to-fine retrieval for camera re-localization

2D-to-3D Localization

Descriptor Matching
ModelsDatePublicationPaperCode
NetVLAD2016CVPRNetvlad: Cnn architecture for weakly supervised place recognition
DELF2017CVPRLarge-scale image retrieval with attentive deep local features
Schonberger et al.2018/06CVPRSemantic Visual Localization
SuperPoint2018CVPRWSuperpoint: Selfsupervised interest point detection and description
NC-Net2018NeurIPSNeighbourhood consensus networks
Sarlin et al.2019/06CVPRFrom Coarse to Fine: Robust Hierarchical Localization at Large Scale
2D3D-MatchNet2019ICRA2d3d-matchnet: learning to match keypoints across 2d image and 3d point cloud
D2-Net2019CVPRD2-net: A trainable cnn for joint description and detection of local features
Speciale et al.2019CVPRPrivacy preserving image-based localization
OOI-Net2019CVPRVisual localization by learning objects-of-interest dense match regression
Camposeco et al.2019CVPRscene compression for visual localization
Cheng et al.2019CVPRCascaded parallel filtering for memory-efficient image-based localization
Taira et al.2019CVPRIs this the right place? geometric-semantic pose verification for indoor visual localization
R2D22019NeurIPSR2d2: Repeatable and reliable detector and descriptor
ASLFeat2020CVPRAslfeat: Learning local features of accurate shape and localization
Scene Coordinate Regression
ModelsDatePublicationPaperCode
DSAC2017/07CVPRDSAC - Differentiable RANSAC for Camera Localization
DSAC++2018/06CVPRLearning less is more-6d camera localization via 3d surface regression
Dense SCR2018/07RSSFull-Frame Scene Coordinate Regression for Image-Based Localization
DSAC++ angle2018/09ECCVScene coordinate regression with angle-based reprojection loss for camera relocalization
Confidence SCR2018/09BMVCScene Coordinate and Correspondence Learning for Image-Based Localization
ESAC2019/10ICCVExpert Sample Consensus Applied to Camera Re-Localization
NG-RANSAC2019/06CVPRNeural-Guided RANSAC: Learning Where to Sample Model Hypotheses
SANet2019/10ICCVSANet: scene agnostic network for camera localization
HSC-Net2020CVPRHierarchical scene coordinate classification and regression for visual localization
KF-Net2020CVPRKfnet: Learning temporal camera relocalization using kalman filtering

3D-to-3D Localization

ModelsDatePublicationPaperCode
LocNet2018IVLocnet: Global localization in 3d point clouds for mobile vehicles
PointNetVLAD2018CVPRPointnetvlad: Deep point cloud based retrieval for large-scale place recognition
Barsan et al.2018CoRLLearning to localize using a lidar intensity map
L3-Net2019CVPRL3-net: Towards learning based lidar localization for autonomous driving
PCAN2019CVPRPCAN: 3D Attention Map Learning Using Contextual Information for Point Cloud Based Retrieval
DeepICP2019CVPRDeepicp: An end-to-end deep neural network for 3d point cloud registration
DCP2019CVPRDeep closest point: Learning representations for point cloud registration
D3Feat2020CVPRD3feat: Joint learning of dense detection and description of 3d local features

SLAM

Local Optimization

ModelsDatePublicationPaperCode
LS-Net2018ECCVLearning to solve nonlinear least squares for monocular stereo
BA-Net2019ICLRBA-Net: Dense bundle adjustment network

Global Optimization

ModelsDatePublicationPaperCode
CNN-SLAM2017CVPRCNN-SLAM: Real-time dense monocular SLAM with learned depth prediction
Li et al.2019ICRAPose graph optimization for unsupervised monocular visual odometry
DeepTAM2020IJCVDeepTAM: Deep Tracking and Mapping with Convolutional Neural Networks
DeepFactors2020RA-LDeepFactors: Real-Time Probabilistic Dense Monocular SLAM

Keyframe and Loop-closure Detection

ModelsDatePublicationPaperCode
Sunderhauf et al.2015RSSPlace recognition with convnet landmarks: Viewpoint-robust, condition-robust, training-free
Gao et al.2017ARUnsupervised learning to detect loops using deep neural networks for visual slam system
Huang et al.2018RSSLightweight unsupervised deep loop closure
Sheng et al.2019ICCVUnsupervised Collaborative Learning of Keyframe Detection and Visual Odometry Towards Monocular Deep SLAM
Memon et al.2020RASLoop closure detection using supervised and unsupervised deep neural networks for monocular slam systems

Uncertainty Estimation

ModelsDatePublicationPaperCode
Kendall et al.2016ICRAModelling uncertainty in deep learning for camera relocalization
Kendall et al.2017NeurIPSWhat Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
VidLoc2017CVPRVidLoc: A Deep Spatio-Temporal Model for 6-DoF Video-Clip Relocalization
Wang et al.2018IJRREnd-to-end, sequenceto-sequence probabilistic visual odometry through deep neural networks
Chen et al.2019TMCDeep neural network based inertial odometry using low-cost inertial measurement units

This list is maintained by Changhao Chen and Bing Wang, Department of Computer Science, University of Oxford.

Please contact them (email: changhao.chen@cs.ox.ac.uk; bing.wang@cs.ox.ac.uk), if you have any question or would like to add your work on this list.