2020年至今-NN SLAM各领域必读的最新研究论文整理分享

    SLAM (simultaneous localization and mapping),也称为CML (Concurrent Mapping and Localization), 即时定位与地图构建，或并发建图与定位。问题可以描述为：将一个机器人放入未知环境中的未知位置，是否有办法让机器人一边移动一边逐步描绘出此环境完全的地图，所谓完全的地图（a consistent map）是指不受障碍行进到房间可进入的每个角落。

    本资源整理了深度学习模型在SLAM（NN SLAM）各个领域的应用，涉及SLAM系统、自监督的SLAM架构、深度估计、视觉里程估计、视觉惯性里程估计、特征表示、摄像机定位、位置识别(环路检测)、地图和地图压缩、路径优化相关的一些近几年最新的论文。

    资源整理自网络，源地址：https://github.com/UltronAI/awesome-nn-slam#slam-system

 

目录

    SLAM架构

    深度估计

    视觉里程估计

    视觉惯性里程估计

    特征表示

    摄像机定位

    位置识别(环路检测)

    地图和地图压缩

    路径优化

 

    SLAM架构

    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

     

    深度估计

    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

     

    视觉里程估计

    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

     

    视觉惯性里程估计

    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

     

    特征表示

    2019

    ·[3DV 2019] SIPs: Succinct Interest Points from Unsupervised Inlierness Probability Learning

    2018

    ·[CVPR 2018 Workshop] SuperPoint: Self-Supervised Interest Point Detection and Description

     

    摄像机定

    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

     

    路径优化

    2019

    ·[ICRA 2019] Pose Graph Optimization for Unsupervised Monocular Visual Odometry 

 

往期精品内容推荐

为AI从业者/研究生/研究员专门定制的全网唯一高端AI训练营

斯坦福、伯克利及MIT联合打-2020年DL/RL必学造6门经典课程-视频资源免费分享

一直在入门从未到精通，成为资深NLP工程师到底需要哪些技能？

2020斯坦福新课-《新冠战疫中的数据科学与AI》视频及ppt分享

少样本学习-7-少样本学习之MAML（模型对抗元学习）

精品干货课程推荐-深度学习、机器学习算法原理、工业实战从入门到精通

《Torch实战教程》免费课程视频及ppt分享

新书-《Bandit问题-基础、算法、理论、应用》分享

Pytorch官方力荐-11月新书《Pytorch深度学习实战指南》pdf及代码分享

机器学习从入门到进阶所需学习资料-包括书、视频、源码

2019年新书-Marc Peter Deisenroth《机器学习基础》-免费分享

中文版《动手学深度学习》9月最新版、代码、ppt、视频分享

2019年Google最新中文版《机器学习速成课程》分享

新书分享-嵌入式深度学习：持续性神经网路算法、结构和电路设计