ECCV20 personal paper list of interest

https://eccv2020.eu/accepted-papers/

1. Robotics
2. Auto. Vehicle.
3. CV
4-. miscs

=================================================================

1. 3D, SfM, SLAM, Localization, Mapping
DeepSFM: Structure From Motion Via Deep Bundle Adjustment
Privacy Preserving Structure-from-Motion
A Consistently Fast and Globally Optimal Solution to the Perspective-n-Point Problem
What Matters in Unsupervised Optical Flow
Coherent full scene 3D reconstruction from a single RGB image
Image-to-Voxel Model Translation for 3D Scene Reconstruction and Segmentation
RobustFusion: Human Volumetric Capture with Data-driven Visual Cues using a RGBD Camera
RANSAC-Flow: generic two-stage image alignment
PointPWC-Net: Cost Volume on Point Clouds for (Self-)Supervised Scene Flow Estimation
DeepGMR: Learning Latent Gaussian Mixture Models for Registration
Shonan Rotation Averaging: Global Optimality by Surfing SO(p)
Consistency Guided Scene Flow Estimation
Image-to-Voxel Model Translation for 3D Scene Reconstruction and Segmentation
Atlas: End-to-End 3D Scene Reconstruction from Posed Images
Beyond Controlled Environments: 3D Camera Re-Localization in Changing Indoor Scenes
Curriculum DeepSDF
Deformable Grid
Relative Pose Estimation of Calibrated Cameras with Known SE(3) Invariants
Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling
??? People as Scene Probes
Scene Scale Estimation from Single Image in the Wild
Pseudo RGB-D for Self-Improving Monocular SLAM and Depth Prediction
??? Contrastive Multiview Coding
Relative pose from deep learned depth and affine correspondences
Rethinking Pseudo-LiDAR Representation
Stochastic Bundle Adjustment for Efficient and Scalable Structure from Motion
??? Polarized optical-flow gyroscope
A Closest Point Proposal for MCMC-based Probabilistic Surface Registration
Continuous Multimodal 6D Camera Relocalization
Structural Deep Metric Learning for Room Layout Estimation
Handcrafted Outlier Detection Revisited
LiteFlowNet3: Resolving Correspondence Ambiguity for More Accurate Optical Flow Estimation
3D-Rotation-Equivariant Quaternion Neural Networks
Unifying Deep Local and Global Features for Image Search
3D Scene Reconstruction from a Single Viewport
LC-VSLAM: Real-time Tracking and Bundle Adjustment in Line-Cloud
Globally Optimal and Efficient Vanishing Point Estimation in Atlanta World
Challenge-Aware RGBT Tracking
Least squares surface reconstruction on arbitrary domains
You Are Here: Geolocation by Embedding Maps and Images
PointTriNet: Learned Triangulation of 3D Point Sets
Deep View Synthesis From Colored 3D PointClouds
*** NeuRoRA: Neural Robust Rotation Averaging
Unsupervised Learning of Optical Flow with Deep Feature Similarity
??? Learning Camera-Aware Noise Models
Iterative Distance-Aware Similarity Matrix Convolution with Mutual-Supervised Point Elimination for Efficient Point Cloud Registration
Simultaneous Detection and Tracking with Motion Modelling for Multiple Object Tracking
Aligning Videos in Space and Time
Efficient Outdoor 3D Point Cloud Semantic Segmentation for Critical Road Objects and Distributed Contexts
Spatial Geometric Reasoning for Room Layout Estimation via Deep Reinforcement Learning
DA4AD: End-to-end Deep Attention Aware Features Aided Visual Localization for Autonomous Driving
LandscapeAR: Large Scale Outdoor Augmented Reality by Matching Photographs with Terrain Models Using Learned Descriptors
Deep Local Shapes: Learning Local SDF Priors for Detailed 3D Reconstruction
Learning Permutation Invariant Representations using Memory Networks
Statistical Outlier Identification in Pose Graphs Using Cycles

1.a point cloud, toy 3D
Intrinsic Point Cloud Interpolation via Dual Latent Space Navigation
Quaternion Equivariant Capsule Networks for 3D Point Clouds
PointMixup: Augmentation for Point Clouds
PointContrast: Unsupervised Pretraining for 3D Point Cloud Understanding
Convolutional Occupancy Networks
Point2Surf: Learning Implicit Surfaces from Point Cloud Patches
Unsupervised 3D Shape Completion in the Wild
Deformation-Aware 3D Shape Embedding and Retrieval
Meshing Point Clouds with Predicted Intrinsic-Extrinsic Ratio Guidance
GRNet: Gridding Residual Network for Dense Point Cloud Completion
Mapping in a Cycle: Sinkhorn Regularized Unsupervised Learning for Point Cloud Shapes
Label-Efficient Learning on Point Clouds using Approximate Convex Decompositions
CN: Channel Normalization in Point Cloud
SPOD: Selective Point Cloud Densification for Better Localization in Point Cloud Object Detection
DPDist : Comparing Point Clouds Using Deep Point Cloud Distance
Weakly Supervised 3D Object Detection from Lidar Point Cloud
Streaming Object Detection for 3-D Point Clouds
PUGeo-Net: A Geometry-centric Network for 3D Point Cloud Upsampling
JSENet: Joint Semantic Segmentation and Edge Detection Network for 3D Point Clouds
DDGCN: A Dynamic Directed Graph Convolutional Network for Action Recognition
A Closer Look at Local Aggregation Operators in Point Cloud Analysis
Discrete Point Flow Networks for Efficient Point Cloud Generation
Unsupervised Learning of Category-Specific Symmetric 3D Keypoints from Point Sets
Efficient Outdoor 3D Point Cloud Semantic Segmentation for Critical Road Objects and Distributed Contexts
3D-CVF: Generating Joint Camera and LiDAR Features Using Cross-View Spatial Feature Fusion for 3D Object Detection
Spatially-Adaptive Convolution for Efficient Point-Cloud Segmentation
Orderly Disorder in Point Cloud Domain
FLOT: Scene Flow Estimation by Learned Optimal Transport on Point Clouds
Volumetric Transformer Networks
Improving Optical Flow on a Pyramid Level
Instance-Aware Embedding for Point Cloud Instance Segmentation

1.b Depth
Du2Net: Learning Depth Estimation from Dual-Cameras and Dual-Pixels
Learning Stereo from Single Images
Mapillary Planet-Scale Depth Dataset
Domain-invariant Stereo Matching Networks
CLIFFNet for Monocular Depth Estimation with Hierarchical Embedding Loss
Occlusion-Aware Depth Estimation with Adaptive Normal Constraints
Pyramid Multi-view Stereo Net with Self-adaptive View Aggregation
Improving Monocular Depth Estimation by Leveraging Structural Awareness and Complementary Datasets
What is Learned in Deep Uncalibrated Photometric Stereo?
Single-Image Depth Prediction Makes Feature Matching Easier
Joint 3D Layout and Depth Prediction from a Single Indoor Panorama Image
HDNet: Human Depth Estimation for Multi-Person Camera-Space Localization
Self-Supervised Monocular Depth Estimation: Solving the Dynamic Object Problem by Semantic Guidance
Depth Estimation by Learning Triangulation and Densification of Sparse Points for Multi-view Stereo
Disambiguating Monocular Depth Estimation with a Single Transient
Sparse-to-Dense Depth Completion Revisited: Sampling Strategy and Graph Construction
Guiding Monocular Depth Estimation Using Depth Attention-Volume
Unsupervised Monocular Depth Estimation for Night-time Images using Adversarial Domain Feature Adaptation

1.c Local Features and matching
Learning and aggregating deep local descriptors for instance-level recognition
Online Invariance Selection for Local Feature Descriptors
S2DNet: Learning accurate correspondences for sparse-to-dense feature matching
Deep Hough Transform for Semantic Line Detection
Shape and Viewpoint without Keypoints
Learning to Compose Hypercolumns for Visual Correspondence
Semantic Line Detection Using Mirror Attention and Comparative Ranking and Matching
Deep Hough-Transform Line Priors

1.d path planning, nav
Learning to plan with uncertain topological maps
Semantic Curiosity for Visual Navigation
Seeing the Un-Scene: Learning Amodal Semantic Maps for Room Navigation
Pillar-based Object Detection for Autonomous Driving
Active Visual Information Gathering for Vision-Language Navigation
Beyond the Nav-Graph: Vision-and-Language Navigation in Continuous Environments
Where to Explore Next? ExHistCNN for History-aware Autonomous 3D Exploration

1.e other sensors
Entropy Minimisation Framework for Event-based Vision Model Estimation
Event-based Asynchronous Sparse Convolutional Networks
??? Learning to See in the Dark with Events
Spike-FlowNet: Event-based Optical Flow Estimation with Energy-Efficient Hybrid Neural Networks

2. Autonomous Vehicle.
Towards Streaming Image Understanding
V2VNet: Vehicle-to-Vehicle Communication for Joint Perception and Prediction
Gen-LaneNet: A Generalized and Scalable Approach for 3D Lane Detection
The Mapillary Traffic Sign Dataset for Detection and Classification on a Global Scale
Ultra Fast Structure-aware Deep Lane Detection
TopoGAN: A Generative Adversarial Approach to Topology-Aware Road Segmentation
Efficient Outdoor 3D Point Cloud Semantic Segmentation for Critical Road Objects and Distributed Contexts
3D-CVF: Generating Joint Camera and LiDAR Features Using Cross-View Spatial Feature Fusion for 3D Object Detection
DA4AD: End-to-end Deep Attention Aware Features Aided Visual Localization for Autonomous Driving
SNE-RoadSeg: Incorporating Surface Normal Information into Semantic Segmentation for Accurate Freespace Detection

3. too CV or too CG or other member-related
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
Rethinking image inpainting via a mutual encoder-decoder with feature equalization
?? Learning to Factorize a City
Guided Deep Decoder: Unsupervised Image Pair Fusion
Optical Flow Distillation: Towards Efficient and Stable Video Style Transfer
Representative Graph Neural Network
When Does Self-supervision Improve Few-shot Learning?
Faster Person Re-Identification
A Tool for Measuring and Mitigating Bias in Visual Datasets
Learnable Cost Volume using the Cayley Representation
Deep Spiking Neural Network: Energy Efficiency Through Time based Coding
Momentum Batch Normalization for Deep Learning with Small Batch Size
Adaptive Offline Quintuplet Loss for Image-text Matching
Actions as Moving Points
L2 Norm: A Generic Visualization Approach for Convolutional Neural Networks
Learning to Learn in a Semi-Supervised Fashion
Dual Adversarial Network for Deep Active Learning
What makes fake images detectable? Understanding properties that generalize
Label-similarity Curriculum Learning
Feature Space Augmentation for Long-Tailed Data

3.b deep priors
LIMP: Learning Latent Shape Representations with Metric Preservation Priors
Filter Style Transfer between Photos
The Hessian Penalty: A Weak Prior for Unsupervised Disentanglement
Deep Image Clustering with Category-Style Representation
Deep Hough-Transform Line Priors
MPCC: Matching Priors and Conditionals for Clustering
Deep Local Shapes: Learning Local SDF Priors for Detailed 3D Reconstruction

3.c open set problem
Hybrid Models for Open Set Recognition
Multiple Class Novelty Detection Under the Data Distribution Shift
Adversarial Continual Learning
Topology-Preserving Class-Incremental Learning
Towards Recognizing Unseen Categories in Unseen Domains
A Boundary Based Out-Of-Distribution Classifier for Generalized Zero-Shot Learning

3.d.1 obj: det, track, pose
Self6D: Self-Supervised Monocular 6D Object Pose Estimation
Corner Proposal Network for Anchor-free, Two-stage Object Detection
Generative Sparse Detection Network for 3D Single-shot Object Detection
Side-Aware Boundary Localization for More Precise Object Detection
Tracking objects as points
PIoU Loss: Towards Accurate Oriented Object Detection in Complex Environments
AABO: Adaptive Anchor Box Optimization for Object Detection via Bayesian Sub-sampling
A General Toolbox for Understanding Errors in Object Detection
Soft Anchor-Point Object Detection
Monocular 3D Object Detection via Feature Domain Adaptation
Point-Set Anchors for Object Detection, Instance Segmentation and Pose Estimation
SPOD: Selective Point Cloud Densification for Better Localization in Point Cloud Object Detection
Cascade Graph Neural Networks for RGB-D Salient Object Detection
PackDet: Packed Long-Head Object Detector
Weakly Supervised 3D Object Detection from Lidar Point Cloud
MimicDet: Bridging the Gap Between One-Stage and Two-Stage Object Detection
OS2D: One-Stage One-Shot Object Detection by Matching Anchor Features
Finding Your (3D) Center: 3D Object Detection Using a Learned Loss
Streaming Object Detection for 3-D Point Clouds
Rotation-robust Intersection over Union for 3D Object Detection
YOLO in the Dark - Domain Adaptation Method for Merging Multiple Models
Pillar-based Object Detection for Autonomous Driving
Dive Deeper Into Box for Object Detection
Probabilistic Anchor Assignment with IoU Prediction for Object Detection
Geometry Constrained Weakly Supervised Object Localization
3D-CVF: Generating Joint Camera and LiDAR Features Using Cross-View Spatial Feature Fusion for 3D Object Detection
Pose Augmentation: Class-agnostic Object Pose Transformation for Object Recognition
Info3D: Representation Learning on 3D Objects using Mutual Information Maximization and Contrastive Learning

3.d.2 seg
SIP: Spatial Information Preservation for Fast Instance Segmentation
Learning with Noisy Class Labels for Instance Segmentation
Boundary-preserving Mask R-CNN
The Devil is in Classification: A Simple Framework for Long-tail Instance Segmentation
Dynamic R-CNN: Towards High Quality Object Detection via Dynamic Training
Learning to Predict Context-adaptive Convolution for Semantic Segmentation
Beyond Monocular Deraining: Paired Rain Removal Networks via Unpaired Semantic Understanding
Efficient Outdoor 3D Point Cloud Semantic Segmentation for Critical Road Objects and Distributed Contexts

3.e net design
Rethinking Bottleneck Structure for Efficient Mobile Network Design
Gabor Layers Enhance Network Robustness
Attentive Normalization
Feature Pyramid Transformer

3.f cls
Distribution-Balanced Loss for Multi-Label Classification in Long-Tailed Datasets
Negative Margin Matters: Understanding Margin in Few-shot Classification
Associative Alignment for Few-shot Image Classification
Learning To Classify Images Without Labels
Rethinking few-shot image classification: a good embedding is all you need?
Topic-aware Multi-Label Classification
Accelerating Deep Learning with Millions of Classes
Pretraining Matters: A Two-Stage Design for Unsupervised Image Classification

4. human pose
End-to-End Estimation of Multi-Person 3D Poses from Multiple Cameras
Appearance Consensus Driven Self-Supervised Human Mesh Recovery
View-Invariant Probabilistic Embedding for Human Pose
Pose2Mesh: Graph Convolutional Network for 3D human Pose and Mesh Recovery from 2D Human Pose
Whole-Body Human Pose Estimation in the Wild
3D Human Shape and Pose from a Single Low-Resolution Image
Explainable Face Recognition
Unsupervised Cross-Modal Alignment For Multi-Person 3D Pose Estimation
Motion Guided 3D Pose Estimation from Video
Unsupervised Human 3D Pose Representation with Viewpoint and Pose Disentanglement
Occlusion-Aware Siamese Network for Human Pose Estimation
NormalGAN: Learning Detailed 3D Human from a Single RGB-D Image -> 저자의 past works

5. metric learning, retrieval
Metric learning: cross-entropy vs. pairwise losses
The Group Loss for Deep Metric Learning
SOLAR: Second-Order Loss and Attention for Image Retrieval
DiVA: Diverse Visual Feature Aggregation for Deep Metric Learning
Deep Credible Metric Learning for Unsupervised Domain Adaptation Person Re-identification
Preserving Semantic Neighborhoods for Robust Cross-modal Retrieval
FeatMatch: Feature-Based Augmentation for Semi-Supervised Learning
Spherical Feature Transform for Deep Metric Learning
Deep Hashing with Active Pairwise Supervision
A Metric Learning Reality Check
A Simple and Effective Framework for Pairwise Deep Metric Learning