SLAM (simultaneous localization and mapping),也稱為CML (Concurrent Mapping and Localization), 即時定位與地圖構建，或並發建圖與定位。問題可以描述為：將一個機器人放入未知環境中的未知位置，是否有辦法讓機器人一邊移動一邊逐步描繪出此環境完全的地圖，所謂完全的地圖（a consistent map）是指不受障礙行進到房間可進入的每個角落。
本資源整理了深度學習模型在SLAM（NN SLAM）各個領域的應用，涉及SLAM系統、自監督的SLAM架構、深度估計、視覺裡程估計、視覺慣性裡程估計、特徵表示、攝像機定位、位置識別(環路檢測)、地圖和地圖壓縮、路徑優化相關的一些近幾年最新的論文。
資源整理自網絡，源地址：https://github.com/UltronAI/awesome-nn-slam#slam-system
目錄
SLAM架構
深度估計
視覺裡程估計
視覺慣性裡程估計
特徵表示
攝像機定位
位置識別(環路檢測)
地圖和地圖壓縮
路徑優化
SLAM架構
2019
·[ICRA 2019] GEN-SLAM: Generative Modeling for Monocular Simultaneous Localization and Mapping
2017
·[CVPR 2017] CNN-SLAM: Real-time dense monocular SLAM with learned depth prediction
Self-supervised Structure-from-Motion
2019
·[ICCV 2019] Self-Supervised Monocular Depth Hints
·[ICCV 2019] Depth from Videos in the Wild: Unsupervised Monocular Depth Learning from Unknown Cameras
·[ICCV 2019] Exploiting temporal consistency for real-time video depth estimation
·[ICCV 2019] Digging Into Self-Supervised Monocular Depth Estimation
·[ICCV 2019] Unsupervised High-Resolution Depth Learning From Videos With Dual Networks
·[ICCV 2019] SynDeMo: Synergistic Deep Feature Alignment for Joint Learning of Depth and Ego-Motion
·[ICCV 2019] Enforcing geometric constraints of virtual normal for depth prediction
·[ICCV 2019] Self-supervised Learning with Geometric Constraints in Monocular Video Connecting Flow, Depth, and Camera
·[ICCV 2019] Moving Indoor: Unsupervised Video Depth Learning in Challenging Environments
·[ICCV 2019] Sequential Adversarial Learning for Self-Supervised Deep Visual Odometry
·[CoRL 2019] Two Stream Networks for Self-Supervised Ego-Motion Estimation
·[IROS 2019] Learning Residual Flow as Dynamic Motion from Stereo Videos
·[NeurIPS 2019] Unsupervised Scale-consistent Depth and Ego-motion Learning from Monocular Video
·[ICRA 2019] Unsupervised Learning of Monocular Depth and Ego-Motion Using Multiple Masks
·[ICRA 2019] GANVO - Unsupervised Deep Monocular Visual Odometry and Depth Estimation with Generative Adversarial Networks
·[CVPR 2019] Competitive Collaboration: Joint Unsupervised Learning of Depth, Camera Motion, Optical Flow and Motion Segmentation
·[CVPR 2019] UnOS: Unified Unsupervised Optical-flow and Stereo-depth Estimation by Watching Videos
·[AAAI 2019] Depth Prediction Without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videos
·[3DV 2019] Structured Coupled Generative Adversarial Networks for Unsupervised Monocular Depth Estimation
·[Arxiv 2019] Flow-Motion and Depth Network for Monocular Stereo and Beyond
2018
·[ECCV 2018] DF-Net: Unsupervised Joint Learning of Depth and Flow using Cross-Task Consistency
·[CVPR 2018] Unsupervised Learning of Monocular Depth Estimation and Visual Odometry with Deep Feature Reconstruction
·[CVPR 2018] GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose
·[IROS 2018] UnDEMoN: Unsupervised Deep Network for Depth and Ego-Motion Estimation
2017
·[CVPR 2017] Unsupervised Learning of Depth and Ego-Motion from Video
深度估計
2019
·[ICCV 2019] How do neural networks see depth in single images?
·[ICCV 2019] Visualization of Convolutional Neural Networks for Monocular Depth Estimation
·[ICCV 2019] Enforcing geometric constraints of virtual normal for depth prediction
·[TPAMI 2019] Progressive Fusion for Unsupervised Binocular Depth Estimation using Cycled Networks
·[CVPR 2019] Student Becoming the Master: Knowledge Amalgamation for Joint Scene Parsing, Depth Estimation, and More
·[CVPR 2019] Learning Monocular Depth Estimation Infusing Traditional Stereo Knowledge
·[CVPR 2019] CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth
·[CVPR 2019] Veritatem Dies Aperit-Temporally Consistent Depth Prediction Enabled by a Multi-Task Geometric and Semantic Scene Understanding Approach
·[CVPR 2019] Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving
·[CVPR 2019] Bilateral Cyclic Constraint and Adaptive Regularization for Unsupervised Monocular Depth Prediction
·[CVPR 2019] Towards Scene Understanding: Unsupervised Monocular Depth Estimation with Semantic-aware Representation
·[CVPR 2019] Geometry-Aware Symmetric Domain Adaptation for Monocular Depth Estimation
·[CVPR 2019] Recurrent MVSNet for High-resolution Multi-view Stereo Depth Inference
·[CVPR 2019] Refine and Distill: Exploiting Cycle-Inconsistency and Knowledge Distillation for Unsupervised Monocular Depth Estimation
·[CVPR 2019] Neural RGB->D Sensing: Depth and Uncertainty from a Video Camera
·[Arxiv 2019] Attention-based Context Aggregation Network for Monocular Depth Estimation
2018
·[ICRA 2018] Just-in-Time Reconstruction: Inpainting Sparse Maps Using Single View Depth Predictors as Priors
·[CVPR 2018] Learning for Disparity Estimation through Feature Constancy
·[CVPR 2018] Deep Ordinal Regression Network for Monocular Depth Estimation
·[CVPR 2018] Learning Depth from Monocular Videos using Direct Methods
·[CVPR 2018] Structured Attention Guided Convolutional Neural Fields for Monocular Depth Estimation
·[CVPR 2018 Workshop] On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach
·[ECCV 2018] Learning Monocular Depth by Distilling Cross-domain Stereo Networks
·[ECCV 2018] Look Deeper into Depth: Monocular Depth Estimation with Semantic Booster and Attention-Driven Loss
2017
·[ICCV 2017] End-to-End Learning of Geometry and Context for Deep Stereo Regression
·[CVPR 2017] Unsupervised Monocular Depth Estimation with Left-Right Consistency
2016 and before
·[ECCV 2016] Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue
·[NeurIPS 2014] Depth Map Prediction from a Single Image Using a Multi-scale Deep Network
視覺裡程估計
2020
·[ICRA 2020] Visual Odometry Revisited: What Should Be Learnt?
2019
·[CVPR 2019] MagicVO: End-to-End Monocular Visual Odometry through Deep Bi-directional Recurrent Convolutional Neural Network
·[CVPR 2019] Understanding the Limitations of CNN-based Absolute Camera Pose Regression
·[CVPR 2019] Beyond Tracking: Selecting Memory and Refining Poses for Deep Visual Odometry
·[ICRA 2019] Learning Monocular Visual Odometry through Geometry-Aware Curriculum Learning
2018
·[ICRA 2018] Deep Auxiliary Learning for Visual Localization and Odometry
·[ICRA 2018] UnDeepVO: Monocular Visual Odometry through Unsupervised Deep Learning
·[CVPR 2018 Workshop] Geometric Consistency for Self-Supervised End-to-End Visual Odometry
·[IJRR 2018] End-to-end, sequence-to-sequence probabilistic visual odometry through deep neural networks
2017
·[ICRA 2017] DeepVO: Towards end-to-end visual odometry with deep Recurrent Convolutional Neural Networks
·[IROS 2017] Deep regression for monocular camera-based 6-DoF global localization in outdoor environments
視覺慣性裡程估計
2019
·[CVPR 2019] Selective Sensor Fusion for Neural Visual-Inertial Odometry
2018
·[IROS 2018] Vision-Aided Absolute Trajectory Estimation Using an Unsupervised Deep Network with Online
2017
·[AAAI 2017] VINet: Visual-Inertial Odometry as a Sequence-to-Sequence Learning ProblemError Correction
特徵表示
2019
·[3DV 2019] SIPs: Succinct Interest Points from Unsupervised Inlierness Probability Learning
2018
·[CVPR 2018 Workshop] SuperPoint: Self-Supervised Interest Point Detection and Description
攝像機定
2019
·[Arxiv 2019] AtLoc: Attention Guided Camera Localization
·[Arxiv 2019] Hierarchical Joint Scene Coordinate Classification and Regression for Visual Localization
2018
·[ICRA 2018] Deep Auxiliary Learning for Visual Localization and Odometry
2017
·[IROS 2017] Deep regression for monocular camera-based 6-DoF global localization in outdoor environments
2016 and before
·[ICCV 2015] PoseNet: A Convolutional Network for Real-Time 6-DOF Camera Relocalization
Place Recognition (Loop Detection)
2016 and before
·[CVPR 2016] NetVLAD: CNN Architecture for Weakly Supervised Place Recognition
路徑優化
2019
·[ICRA 2019] Pose Graph Optimization for Unsupervised Monocular Visual Odometry