Paper notes

This repository contains my paper reading notes on deep learning and machine learning. It is inspired by Denny Britz and Daniel Takeshi. A minimalistic webpage generated with Github io can be found here.

About me

My name is Patrick Langechuan Liu. After about a decade of education and research in physics, I found my passion in deep learning and autonomous driving. Currently I am leading the development of perception features at Xpeng Motors, a fast growing autonomous driving company.

What to read

Where to start?

If you are new to deep learning in computer vision and don't know where to start, I suggest you spend your first month or so dive deep into this list of papers. I did so (see my notes) and it served me well.

Here is a list of trustworthy sources of papers in case I ran out of papers to read.

My Review Posts by Topics

I regularly update my blog in Toward Data Science.

• Monocular BEV Perception with Transformers in Autonomous Driving (related paper notes)
• Illustrated Differences between MLP and Transformers for Tensor Reshaping in Deep Learning
• Monocular 3D Lane Line Detection in Autonomous Driving (related paper notes)
• Deep-Learning based Object detection in Crowded Scenes (related paper notes)
• Monocular Bird’s-Eye-View Semantic Segmentation for Autonomous Driving (related paper notes)
• Deep Learning in Mapping for Autonomous Driving
• Monocular Dynamic Object SLAM in Autonomous Driving
• Monocular 3D Object Detection in Autonomous Driving — A Review
• Self-supervised Keypoint Learning — A Review
• Single Stage Instance Segmentation — A Review
• Self-paced Multitask Learning — A Review
• Convolutional Neural Networks with Heterogeneous Metadata
• Lifting 2D object detection to 3D in autonomous driving
• Multimodal Regression
• Paper Reading in 2019

2023-05

• ProphNet: Efficient Agent-Centric Motion Forecasting with Anchor-Informed Proposals CVPR 2023 [Qcraft, prediction]
• SAM: Segment Anything [FAIR]
• GeoMIM: Towards Better 3D Knowledge Transfer via Masked Image Modeling for Multi-view 3D Understanding
• Motion Prediction using Trajectory Sets and Self-Driving Domain Knowledge [Encode Road requirement to prediction]
• [Hivt: Hierarchical vector transformer for multi-agent motion prediction]
• Transformer Feed-Forward Layers Are Key-Value Memories EMNLP 2021
• BEV-LaneDet: a Simple and Effective 3D Lane Detection Baseline CVPR 2023 [BEVNet]
• Exploring Recurrent Long-term Temporal Fusion for Multi-view 3D Perception [BEVNet, megvii]
• VAD: Vectorized Scene Representation for Efficient Autonomous Driving [Horizon]
• A Simple Attempt for 3D Occupancy Estimation in Autonomous Driving
• BEVPoolv2: A Cutting-edge Implementation of BEVDet Toward Deployment [BEVDet, PhiGent]
• NVRadarNet: Real-Time Radar Obstacle and Free Space Detection for Autonomous Driving
• GraspNet-1Billion: A Large-Scale Benchmark for General Object Grasping CVPR 2020 [Cewu Lu]
• AnyGrasp: Robust and Efficient Grasp Perception in Spatial and Temporal Domains [Cewu Lu]
• SSCNet: Semantic Scene Completion from a Single Depth Image CVPR 2017
• 3D Semantic Scene Completion: a Survey IJCV 2023
• Point Cloud Forecasting as a Proxy for 4D Occupancy Forecasting
• HDGT: Heterogeneous Driving Graph Transformer for Multi-Agent Trajectory Prediction via Scene Encoding
• MTR: Motion Transformer with Global Intention Localization and Local Movement Refinement NeurIPS 2022
• Occupancy Networks: Learning 3D Reconstruction in Function Space CVPR 2019 [Andreas Geiger]
• UVTR: Unifying Voxel-based Representation with Transformer for 3D Object Detection [BEVFusion, Megvii, BEVNet, camera + lidar]
• Don't Use Large Mini-Batches, Use Local SGD ICLR 2020
• Grokking: Generalization beyond Overfitting on small algorithmic datasets
• Progress measures for grokking via mechanistic interpretability
• Understanding deep learning requires rethinking generalization ICLR 2017
• Unifying Grokking and Double DescentDownload
• Deep Interactive Motion Prediction and Planning: Playing Games with Motion Prediction Models L4DC 2022
• Interactive Prediction and Planning for Autonomous Driving: from Algorithms to Fundamental Aspects [PhD thesis of Wei Zhan, 2019]
• Lyft1001: One Thousand and One Hours: Self-driving Motion Prediction Dataset [Lyft Level 5, prediction dataset]

2023-04 (1)

• UniAD: Planning-oriented Autonomous Driving [Notes] [BEV, e2e, Hongyang Li]

2023-03 (5)

• GPT-4 Technical Report [Notes] [OpenAI, GPT]
• OpenOccupancy: A Large Scale Benchmark for Surrounding Semantic Occupancy Perception [Notes] [Occupancy Network, Jiwen Lu]
• VoxFormer: Sparse Voxel Transformer for Camera-based 3D Semantic Scene Completion [Note] [Occupancy Network, Nvidia]
• MonoScene: Monocular 3D Semantic Scene Completion CVPR 2022 [Notes] [Occupancy Network, single cam]
• CoReNet: Coherent 3D scene reconstruction from a single RGB image [Notes] ECCV 2020 oral

2023-02 (4)

• Will we run out of data? An analysis of the limits of scaling datasets in Machine Learning [Notes] [Epoch.ai industry report]
• Codex: Evaluating Large Language Models Trained on Code [Notes] [GPT, OpenAI]
• InstructGPT: Training language models to follow instructions with human feedback [Notes] [GPT, OpenAI]
• TPVFormer: Tri-Perspective View for Vision-Based 3D Semantic Occupancy Prediction [Notes] [Occupancy Network, Jiwen Lu]

2023-01 (2)

• PPGeo: Policy Pre-training for End-to-end Autonomous Driving via Self-supervised Geometric Modeling [Notes] ICLR 2023
• nuPlan: A closed-loop ML-based planning benchmark for autonomous vehicles [Notes]
• CoBEVT: Cooperative Bird's Eye View Semantic Segmentation with Sparse Transformers CoRL 2022
• BEVFormer v2: Adapting Modern Image Backbones to Bird's-Eye-View Recognition via Perspective Supervision [BEVNet, Jifeng Dai]
• Fast-BEV: Towards Real-time On-vehicle Bird’s-Eye View Perception NeurIPS 2022
• Traj++: Human Trajectory Forecasting in Crowds: A Deep Learning Perspective TITS 2021
• Data Driven Prediction Architecture for Autonomous Driving and its Application on Apollo Platform IV 2020 [Baidu]
• THOMAS: Trajectory Heatmap Output with learned Multi-Agent Sampling ICLR 2022
• Learning Lane Graph Representations for Motion Forecasting ECCV 2020 oral
• Identifying Driver Interactions via Conditional Behavior Prediction ICRA 2021 [Waymo]
• Trajectron++: Dynamically-Feasible Trajectory Forecasting With Heterogeneous Data ECCV 2020
• TPNet: Trajectory Proposal Network for Motion Prediction CVPR 2020
• GOHOME: Graph-Oriented Heatmap Output for future Motion Estimation
• PECNet: It Is Not the Journey but the Destination: Endpoint Conditioned Trajectory Prediction ECCV 2020 oral
• From Goals, Waypoints & Paths To Long Term Human Trajectory Forecasting ICCV 2019
• PRECOG: PREdiction Conditioned On Goals in Visual Multi-Agent Settings ICCV 2019
• PiP: Planning-informed Trajectory Prediction for Autonomous Driving ECCV 2020
• MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Prediction CoRL 2019
• LaPred: Lane-Aware Prediction of Multi-Modal Future Trajectories of Dynamic Agents CVPR 2021
• PRIME: Learning to Predict Vehicle Trajectories with Model-based Planning CoRL 2021
• A Flexible and Explainable Vehicle Motion Prediction and Inference Framework Combining Semi-Supervised AOG and ST-LSTM TITS 2020
• Multi-Modal Trajectory Prediction of Surrounding Vehicles with Maneuver based LSTMs IV 2018 [Trivedi]
• HYPER: Learned Hybrid Trajectory Prediction via Factored Inference and Adaptive Sampling ICRA 2022
• Trajectory Prediction with Linguistic Representations ICRA 2022
• What-If Motion Prediction for Autonomous Driving
• End-to-end Contextual Perception and Prediction with Interaction Transformer IROS 2020 [Auxiliary collision loss, scene compliant pred]
• SafeCritic: Collision-Aware Trajectory Prediction BMVC 2019 [IRL, scene compliant pred]
• Large Scale Interactive Motion Forecasting for Autonomous Driving: The Waymo Open Motion Dataset ICCV 2021 [Waymo]
• Interaction-Based Trajectory Prediction Over a Hybrid Traffic Graph IROS 2020
• Joint Interaction and Trajectory Prediction for Autonomous Driving using Graph Neural Networks NeurIPS 2019 workshop
• Fast Risk Assessment for Autonomous Vehicles Using Learned Models of Agent Futures Robotics: science and systems 2020
• Monocular 3D Object Detection: An Extrinsic Parameter Free Approach CVPR 2021 [PJLab]
• UniFormer: Unified Multi-view Fusion Transformer for Spatial-Temporal Representation in Bird's-Eye-View [BEVFormer, BEVNet, Temporal]
• GitNet: geometric prior-baesd transformation for birds yee view segmentation
• WBF: weighted box fusion: ensembling boxes from differnt object detection modules
• NNI: auto parameter finding algorithm
• BEVFormer++: Improving BEVFormer for 3D Camera-only Object Detection [Waymo open dataset challenge 1st place in mono3d]
• LET-3D-AP: Longitudinal Error Tolerant 3D Average Precision for Camera-Only 3D Detection [Waymo open dataset challenge official metric]
• High-Level Interpretation of Urban Road Maps Fusing Deep Learning-Based Pixelwise Scene Segmentation and Digital Navigation Maps Journal of Advanced Transportation 2018
• A Hybrid Vision-Map Method for Urban Road Detection Journal of Advanced Transportation 2017
• Terminology and Analysis of Map Deviations in Urban Domains: Towards Dependability for HD Maps in Automated Vehicles IV 2020
• TIME WILL TELL: NEW OUTLOOKS AND A BASELINE FOR TEMPORAL MULTI-VIEW 3D OBJECT DETECTION
• Conditional DETR for Fast Training Convergence ICCV 2021
• DAB-DETR: Dynamic Anchor Boxes are Better Queries for DETR ICLR 2022
• DN-DETR: Accelerate DETR Training by Introducing Query DeNoising CVPR 2022
• DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection
• Trajectory Forecasting from Detection with Uncertainty-Aware Motion Encoding [Ouyang Wanli]
• Vision-based Uneven BEV Representation Learning with Polar Rasterization and Surface Estimation [BEVNet, polar]
• MUTR3D: A Multi-camera Tracking Framework via 3D-to-2D Queries [BEVNet, tracking] CVPR 2022 workshop [Hang Zhao]
• ST-P3: End-to-end Vision-based Autonomous Driving via Spatial-Temporal Feature Learning ECCV 2022 [Hongyang Li]
• GKT: Efficient and Robust 2D-to-BEV Representation Learning via Geometry-guided Kernel Transformer [BEVNet, Horizon]
• SiamRPN: High Performance Visual Tracking with Siamese Region Proposal Network CVPR 2018
• TPLR: Topology Preserving Local Road Network Estimation from Single Onboard Camera Image CVPR 2022 [STSU, Luc Van Gool]
• LaRa: Latents and Rays for Multi-Camera Bird's-Eye-View Semantic Segmentation [Valeo, BEVNet, polar]
• PolarDETR: Polar Parametrization for Vision-based Surround-View 3D Detection [BEVNet]
• Exploring Geometric Consistency for Monocular 3D Object Detection CVPR 2022
• ImVoxelNet: Image to Voxels Projection for Monocular and Multi-View General-Purpose 3D Object Detection WACV 2022 [mono3D]
• Learning to Predict 3D Lane Shape and Camera Pose from a Single Image via Geometry Constraints AAAI 2022
• Detecting Lane and Road Markings at A Distance with Perspective Transformer Layers ICICN 2021 [BEVNet, lane line]
• Unsupervised Labeled Lane Markers Using Maps ICCV 2019 workshop [Bosch, 2D lane line]
• M3DeTR: Multi-representation, Multi-scale, Mutual-relation 3D Object Detection with Transformers [Lidar detection, Waymo open dataset] WACV 2022
• K-Lane: Lidar Lane Dataset and Benchmark for Urban Roads and Highways [lane line dataset]
• Robust Monocular 3D Lane Detection With Dual Attention ICIP 2021
• OcclusionFusion: Occlusion-aware Motion Estimation for Real-time Dynamic 3D Reconstruction CVPR 2022
• MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer ICLR 2022 [lightweight Transformers]
• XFormer: Lightweight Vision Transformer with Cross Feature Attention [Samsung]
• CenterFormer: Center-based Transformer for 3D Object Detection ECCV 2022 oral [TuSimple]
• LidarMultiNet: Towards a Unified Multi-task Network for LiDAR Perception [2022 Waymo Open Dataset, TuSimple]
• MTRA: 1st Place Solution for 2022 Waymo Open Dataset Challenge - Motion Prediction [Waymo open dataset challenge 1st place in motion prediction]
• BEVSegFormer: Bird's Eye View Semantic Segmentation From Arbitrary Camera Rigs [BEVNet]
• Panoptic SegFormer: Delving Deeper into Panoptic Segmentation with Transformers CVPR 2022 [nVidia]
• Efficiently Identifying Task Groupings for Multi-Task Learning NeurIPS 2021 spotlight [MTL]
• Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time [Google, Golden Backbone]
• "The Pedestrian next to the Lamppost" Adaptive Object Graphs for Better Instantaneous Mapping CVPR 2022
• GitNet: Geometric Prior-based Transformation for Birds-Eye-View Segmentation [BEVNet, Baidu]
• FUTR3D: A Unified Sensor Fusion Framework for 3D Detection [Hang Zhao]
• GitNet: Geometric Prior-based Transformation for Birds-Eye-View Segmentation [BEVNet]
• MonoFormer: Towards Generalization of self-supervised monocular depth estimation with Transformers [monodepth]
• Time3D: End-to-End Joint Monocular 3D Object Detection and Tracking for Autonomous Driving
• cosFormer: Rethinking Softmax in Attention ICLR 2022
• StretchBEV: Stretching Future Instance Prediction Spatially and Temporally [BEVNet, prediction]
• Scene Representation in Bird’s-Eye View from Surrounding Cameras with Transformers [BEVNet, LLD] CVPR 2022 workshop
• Multi-Frame Self-Supervised Depth with Transformers CVPR 2022
• It's About Time: Analog Clock Reading in the Wild CVPR 2022 [Andrew Zisserman]
• SurroundDepth: Entangling Surrounding Views for Self-Supervised Multi-Camera Depth Estimation [Jiwen Lu]
• ONCE-3DLanes: Building Monocular 3D Lane Detection CVPR 2022
• K-Lane: Lidar Lane Dataset and Benchmark for Urban Roads and Highways CVPR 2022 workshop [3D LLD]
• Multi-modal 3D Human Pose Estimation with 2D Weak Supervision in Autonomous Driving CVPR 2022 workshop
• A Simple Baseline for BEV Perception Without LiDAR [TRI, BEVNet, vision+radar]
• Reconstruct from Top View: A 3D Lane Detection Approach based on Geometry Structure Prior CVPR 2022 workshop
• RIDDLE: Lidar Data Compression with Range Image Deep Delta Encoding CVPR 2022 [Waymo, Charles Qi]
• Occupancy Flow Fields for Motion Forecasting in Autonomous Driving RAL 2022 [Waymo occupancy flow challenge]
• Safe Local Motion Planning with Self-Supervised Freespace Forecasting CVPR 2021
• 数据闭环的核心 - Auto-labeling 方案分享
• K-Lane: Lidar Lane Dataset and Benchmark for Urban Roads and Highways
• LETR: Line Segment Detection Using Transformers without Edges CVPR 2021 oral
• HDMapGen: A Hierarchical Graph Generative Model of High Definition Maps CVPR 2021 [HD mapping]
• SketchRNN: A Neural Representation of Sketch Drawings [David Ha]
• PolyGen: An Autoregressive Generative Model of 3D Meshes ICML 2020
• SOLQ: Segmenting Objects by Learning Queries NeurlPS 2021 [Megvii, end-to-end, instance segmentation]
• MonoViT: Self-Supervised Monocular Depth Estimation with a Vision Transformer 3DV 2022
• MVSTER: Epipolar Transformer for Efficient Multi-View Stereo ECCV 2022
• MOVEDepth: Crafting Monocular Cues and Velocity Guidance for Self-Supervised Multi-Frame Depth Learning [MVS + monodepth]
• SurroundDepth: Entangling Surrounding Views for Self-Supervised Multi-Camera Depth Estimation
• Scene Transformer: A unified architecture for predicting multiple agent trajectories [prediction, Waymo] ICLR 2022
• SSIA: Monocular Depth Estimation with Self-supervised Instance Adaptation [VGG team, TTR, test time refinement, CVD]
• CoMoDA: Continuous Monocular Depth Adaptation Using Past Experiences WACV 2021
• MonoRec: Semi-supervised dense reconstruction in dynamic environments from a single moving camera CVPR 2021 [Daniel Cremmers]
• Plenoxels: Radiance Fields without Neural Networks
• Lidar with Velocity: Motion Distortion Correction of Point Clouds from Oscillating Scanning Lidars [Livox, ISEE]
• NWD: A Normalized Gaussian Wasserstein Distance for Tiny Object Detection
• Towards Optimal Strategies for Training Self-Driving Perception Models in Simulation NeurIPS 2021 [Sanja Fidler]
• Insta-DM: Learning Monocular Depth in Dynamic Scenes via Instance-Aware Projection Consistency AAAI 2021
• Instance-wise Depth and Motion Learning from Monocular Videos NeurIPS 2020 workshop [website]
• NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis ECCV 2020 oral
• BARF: Bundle-Adjusting Neural Radiance Fields ICCV 2021 oral
• NerfingMVS: Guided Optimization of Neural Radiance Fields for Indoor Multi-view Stereo ICCV 2021 oral
• Transfuser: Multi-Modal Fusion Transformer for End-to-End Autonomous Driving CVPR 2021
• YOLinO: Generic Single Shot Polyline Detection in Real Time ICCV 2021 workshop [lld]
• MonoRCNN: Geometry-based Distance Decomposition for Monocular 3D Object Detection ICCV 2021
• MonoCInIS: Camera Independent Monocular 3D Object Detection using Instance Segmentation ICCV 2021 workshop
• PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection CVPR 2020 [Waymo challenge 2nd place]
• Geometry-based Distance Decomposition for Monocular 3D Object Detection ICCV 2021 [mono3D]
• Offboard 3D Object Detection from Point Cloud Sequences CVPR 2021 [Charles Qi]
• FreeAnchor: Learning to Match Anchors for Visual Object Detection NeurIPS 2019
• AutoAssign: Differentiable Label Assignment for Dense Object Detection
• Probabilistic Anchor Assignment with IoU Prediction for Object Detection ECCV 2020
• FOVEA: Foveated Image Magnification for Autonomous Navigation ICCV 2021 [Argo]
• PifPaf: Composite Fields for Human Pose Estimation CVPR 2019
• Monocular 3D Localization of Vehicles in Road Scenes ICCV 2021 workshop [mono3D, tracking]
• TransformerFusion: Monocular RGB Scene Reconstruction using Transformers
• Conditional DETR for Fast Training Convergence
• Anchor DETR: Query Design for Transformer-Based Detector [megvii]
• PGD: Probabilistic and Geometric Depth: Detecting Objects in Perspective CoRL 2021
• Adaptive Wing Loss for Robust Face Alignment via Heatmap Regression
• What Makes for End-to-End Object Detection? PMLR 2021
• Instances as Queries ICCV 2021 [instance segmentation]
• One Million Scenes for Autonomous Driving: ONCE Dataset [Huawei]
• NVS-MonoDepth: Improving Monocular Depth Prediction with Novel View Synthesis 3DV 2021
• Is 2D Heatmap Representation Even Necessary for Human Pose Estimation?
• Topology Preserving Local Road Network Estimation from Single Onboard Camera Image [BEVNet, Luc Van Gool]

2022-11 (1)

• M2I: From Factored Marginal Trajectory Prediction to Interactive Prediction [Notes] CVPR 2022

2022-10 (1)

• Delving into the Devils of Bird's-eye-view Perception: A Review, Evaluation and Recipe [Notes] [PJLab]

2022-09 (3)

• ViP3D: End-to-end Visual Trajectory Prediction via 3D Agent Queries [Notes] [BEV, perception + prediction, Hang Zhao]
• MapTR: Structured Modeling and Learning for Online Vectorized HD Map Construction [Notes] [Horizon, BEVNet]
• StopNet: Scalable Trajectory and Occupancy Prediction for Urban Autonomous Driving ICRA 2022
• MOTR: End-to-End Multiple-Object Tracking with Transformer ECCV 2022 [Megvii, MOT]
• Anchor DETR: Query Design for Transformer-Based Object Detection [Notes] AAAI 2022 [Megvii]

2022-08 (1)

• HOME: Heatmap Output for future Motion Estimation [Notes] ITSC 2021 [behavior prediction, Huawei Paris]

2022-07 (8)

• PersFormer: 3D Lane Detection via Perspective Transformer and the OpenLane Benchmark [Notes] [BEVNet, lane line]
• VectorMapNet: End-to-end Vectorized HD Map Learning [Notes] [BEVNet, LLD, Hang Zhao]
• PETR: Position Embedding Transformation for Multi-View 3D Object Detection [Notes] ECCV 2022 [BEVNet]
• PETRv2: A Unified Framework for 3D Perception from Multi-Camera Images [Notes] [BEVNet, MegVii]
• M^2BEV: Multi-Camera Joint 3D Detection and Segmentation with Unified Birds-Eye View Representation [Notes] [BEVNet, nvidia]
• BEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object Detection [Notes] [BEVNet, NuScenes SOTA, Megvii]
• CVT: Cross-view Transformers for real-time Map-view Semantic Segmentation [Notes] CVPR 2022 oral [UTAustin, Philipp]
• Wayformer: Motion Forecasting via Simple & Efficient Attention Networks [Notes] [Behavior prediction, Waymo]

2022-06 (3)

• BEVDet4D: Exploit Temporal Cues in Multi-camera 3D Object Detection [Notes] [BEVNet]
• BEVerse: Unified Perception and Prediction in Birds-Eye-View for Vision-Centric Autonomous Driving [Notes] [Jiwen Lu, BEVNet, perception + prediction]
• BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representation [Notes] [BEVNet, Han Song]

2022-03 (1)

• BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers [Notes] [BEVNet, Hongyang Li, Jifeng Dai]

2022-02 (1)

• TNT: Target-driveN Trajectory Prediction [Notes] CoRL 2020 [prediction, Waymo, Hang Zhao]
• DenseTNT: End-to-end Trajectory Prediction from Dense Goal Sets [Notes] ICCV 2021 [prediction, Waymo, 1st place winner WOMD]

2022-01 (1)

• Manydepth: The Temporal Opportunist: Self-Supervised Multi-Frame Monocular Depth [Notes] CVPR 2021 [monodepth, Niantic]
• DEKR: Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression [Notes] CVPR 2021

2021-12 (5)

• BN-FFN-BN: Leveraging Batch Normalization for Vision Transformers [Notes] ICCVW 2021 [BN, transformers]
• PowerNorm: Rethinking Batch Normalization in Transformers [Notes] ICML 2020 [BN, transformers]
• MultiPath++: Efficient Information Fusion and Trajectory Aggregation for Behavior Prediction [Notes] ICRA 2022 [Waymo, behavior prediction]
• BEVDet: High-performance Multi-camera 3D Object Detection in Bird-Eye-View [Notes]
• Translating Images into Maps [Notes] ICRA 2022 [BEVNet, transformers]

2021-11 (4)

• DETR3D: 3D Object Detection from Multi-view Images via 3D-to-2D Queries [Notes] CoRL 2021 [BEVNet, transformers]
• Robust-CVD: Robust Consistent Video Depth Estimation CVPR 2021 oral [website]
• MAE: Masked Autoencoders Are Scalable Vision Learners [Notes] [Kaiming He, unsupervised learning]
• SimMIM: A Simple Framework for Masked Image Modeling [Notes] [MSRA, unsupervised learning, MAE]
• iBOT: Image BERT Pre-Training with Online Tokenizer

2021-10 (3)

• STSU: Structured Bird's-Eye-View Traffic Scene Understanding from Onboard Images [Notes] ICCV 2021 [BEV feat stitching, Luc Van Gool]
• PanopticBEV: Bird's-Eye-View Panoptic Segmentation Using Monocular Frontal View Images [Notes] RAL 2022 [BEVNet, vertical/horizontal features]
• NEAT: Neural Attention Fields for End-to-End Autonomous Driving [Notes] ICCV 2021 [supplementary] [BEVNet]

2021-09 (11)

• DD3D: Is Pseudo-Lidar needed for Monocular 3D Object detection? [Notes] ICCV 2021 [mono3D, Toyota]
• EfficientDet: Scalable and Efficient Object Detection [Notes] CVPR 2020 [BiFPN, Tesla AI day]
• PnPNet: End-to-End Perception and Prediction with Tracking in the Loop [Notes] CVPR 2020 [Uber ATG]
• MP3: A Unified Model to Map, Perceive, Predict and Plan [Notes] CVPR 2021 [Uber, planning]
• BEV-Net: Assessing Social Distancing Compliance by Joint People Localization and Geometric Reasoning [Notes] ICCV 2021 [BEVNet, surveillance]
• LiDAR R-CNN: An Efficient and Universal 3D Object Detector [Notes] CVPR 2021 [TuSimple, Naiyan Wang]
• Corner Cases for Visual Perception in Automated Driving: Some Guidance on Detection Approaches [Notes] [corner cases]
• Systematization of Corner Cases for Visual Perception in Automated Driving [Notes] IV 2020 [corner cases]
• An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Driving [Notes] IV 2021 [corner cases]
• PYVA: Projecting Your View Attentively: Monocular Road Scene Layout Estimation via Cross-view Transformation [Notes] CVPR 2021 [Supplementary] [BEVNet]
• YOLOF: You Only Look One-level Feature [Notes] CVPR 2021 [megvii]
• Perceiving Humans: from Monocular 3D Localization to Social Distancing [Notes] TITS 2021 [monoloco++]

2021-08 (11)

• EgoNet: Exploring Intermediate Representation for Monocular Vehicle Pose Estimation [Notes] CVPR 2021 [mono3D]
• MonoEF: Monocular 3D Object Detection: An Extrinsic Parameter Free Approach [Notes] CVPR 2021 [mono3D]
• GAC: Ground-aware Monocular 3D Object Detection for Autonomous Driving [Notes] RAL 2021 [mono3D]
• FCOS3D: Fully Convolutional One-Stage Monocular 3D Object Detection [Notes] NeurIPS 2020 [mono3D, senseTime]
• GUPNet: Geometry Uncertainty Projection Network for Monocular 3D Object Detection [Notes] ICCV 2021 [mono3D, Wanli Ouyang]
• DARTS: Differentiable Architecture Search [Notes] ICLR 2019 [VGG author]
• FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search [Notes] CVPR 20219 [DARTS]
• FBNetV2: Differentiable Neural Architecture Search for Spatial and Channel Dimensions CVPR 2020
• FBNetV3: Joint Architecture-Recipe Search using Predictor Pretraining CVPR 2021
• Perceiver: General Perception with Iterative Attention [Notes] ICML 2021 [transformers, multimodal]
• Perceiver IO: A General Architecture for Structured Inputs & Outputs [Notes]
• PillarMotion: Self-Supervised Pillar Motion Learning for Autonomous Driving [Notes] CVPR 2021 [Qcraft, Alan Yuille]
• SimTrack: Exploring Simple 3D Multi-Object Tracking for Autonomous Driving [Notes] ICCV 2019 [QCraft, Alan Yuille]

2021-07 (1)

• HDMapNet: An Online HD Map Construction and Evaluation Framework [Notes] CVPR 2021 workshop [youtube video only, Li Auto]

2021-06 (2)

• FIERY: Future Instance Prediction in Bird's-Eye View from Surround Monocular Cameras [Notes] ICCV 2021 [BEVNet, perception + prediction]
• Baidu's CNN seg [Notes]

2021-04 (5)

• Rethinking the Heatmap Regression for Bottom-up Human Pose Estimation [Notes] CVPR 2021 [megvii]
• CrowdPose: Efficient Crowded Scenes Pose Estimation and A New Benchmark CVPR 2019
• The Overlooked Elephant of Object Detection: Open Set WACV 2021
• Class-Agnostic Object Detection WACV 2021
• OWOD: Towards Open World Object Detection [Notes] CVPR 2021 oral
• FsDet: Frustratingly Simple Few-Shot Object Detection ICML 2020
• MonoFlex: Objects are Different: Flexible Monocular 3D Object Detection [Notes] CVPR 2021 [mono3D, Jiwen Lu, cropped]
• monoDLE: Delving into Localization Errors for Monocular 3D Object Detection [Notes] CVPR 2021 [mono3D]
• Exploring 2D Data Augmentation for 3D Monocular Object Detection
• OCM3D: Object-Centric Monocular 3D Object Detection [mono3D]
• FSM: Full Surround Monodepth from Multiple Cameras [Notes] ICRA 2021 [monodepth, Xnet]

2021-03 (4)

• CaDDN: Categorical Depth Distribution Network for Monocular 3D Object Detection [Notes] CVPR 2021 oral [mono3D, BEVNet]
• DSNT: Numerical Coordinate Regression with Convolutional Neural Networks [Notes] [differentiable spatial to numerical transform]
• Soft-Argmax: Human pose regression by combining indirect part detection and contextual information
• INSTA-YOLO: Real-Time Instance Segmentation [Notes] ICML workshop 2020 [single stage instance segmentation]
• CenterNet2: Probabilistic two-stage detection [Notes] [CenterNet, two-stage]

2021-01 (7)

• Confluence: A Robust Non-IoU Alternative to Non-Maxima Suppression in Object Detection [Notes] [NMS]
• BoxInst: High-Performance Instance Segmentation with Box Annotations [Notes] CVPR 2021 [Chunhua Shen, Tian Zhi]
• 3DSSD: Point-based 3D Single Stage Object Detector [Notes] CVPR 2020
• RepVGG: Making VGG-style ConvNets Great Again [Notes] [Megvii, Xiangyu Zhang, ACNet]
• ACNet: Strengthening the Kernel Skeletons for Powerful CNN via Asymmetric Convolution Blocks [Notes] ICCV 2019
• BEV-Feat-Stitching: Understanding Bird's-Eye View Semantic HD-Maps Using an Onboard Monocular Camera [Notes] [BEVNet, mono3D, Luc Van Gool]
• PSS: Object Detection Made Simpler by Eliminating Heuristic NMS [Notes] [Transformer, DETR]

2020-12 (17)

• DeFCN: End-to-End Object Detection with Fully Convolutional Network [Notes] [Transformer, DETR]
• OneNet: End-to-End One-Stage Object Detection by Classification Cost [Notes] [Transformer, DETR]
• Traffic Light Mapping, Localization, and State Detection for Autonomous Vehicles [Notes] ICRA 2011 [traffic light, Sebastian Thrun]
• Towards lifelong feature-based mapping in semi-static environments [Notes] ICRA 2016
• How to Keep HD Maps for Automated Driving Up To Date [Notes] ICRA 2020 [BMW]
• Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection [Notes] CVPR 2021 [focal loss]
• Visual SLAM for Automated Driving: Exploring the Applications of Deep Learning [Notes] CVPR 2018 workshop
• Centroid Voting: Object-Aware Centroid Voting for Monocular 3D Object Detection [Notes] IROS 2020 [mono3D, geometry + appearance = distance]
• Monocular 3D Object Detection in Cylindrical Images from Fisheye Cameras [Notes] [GM Israel, mono3D]
• DeepPS: Vision-Based Parking-Slot Detection: A DCNN-Based Approach and a Large-Scale Benchmark Dataset TIP 2018 [Parking slot detection, PS2.0 dataset]
• PSDet: Efficient and Universal Parking Slot Detection [Notes] IV 2020 [Zongmu, Parking slot detection]
• PatDNN: Achieving Real-Time DNN Execution on Mobile Devices with Pattern-based Weight Pruning [Notes] ASPLOS 2020 [pruning]
• Scaled-YOLOv4: Scaling Cross Stage Partial Network [Notes] [yolo]
• Yolov5 by Ultralytics [Notes] [yolo, spatial2channel]
• PP-YOLO: An Effective and Efficient Implementation of Object Detector [Notes] [yolo, paddle-paddle, baidu]
• PointPainting: Sequential Fusion for 3D Object Detection [Notes] [nuscenece]
• MotionNet: Joint Perception and Motion Prediction for Autonomous Driving Based on Bird's Eye View Maps [Notes] CVPR 2020 [Unseen moving objects, BEV]
• Locating Objects Without Bounding Boxes [Notes] CVPR 2019 [weighted Haussdorf distance, NMS-free]

2020-11 (18)

• TSP: Rethinking Transformer-based Set Prediction for Object Detection [Notes] ICCV 2021 [DETR, transformers, Kris Kitani]
• Sparse R-CNN: End-to-End Object Detection with Learnable Proposals [Notes] CVPR 2020 [DETR, Transformer]
• Unsupervised Monocular Depth Learning in Dynamic Scenes [Notes] CoRL 2020 [LearnK improved ver, Google]
• MoNet3D: Towards Accurate Monocular 3D Object Localization in Real Time [Notes] ICML 2020 [Mono3D, pairwise relationship]
• Argoverse: 3D Tracking and Forecasting with Rich Maps [Notes] CVPR 2019 [HD maps, dataset, CV lidar]
• The H3D Dataset for Full-Surround 3D Multi-Object Detection and Tracking in Crowded Urban Scenes [Notes] ICRA 2019
• Cityscapes 3D: Dataset and Benchmark for 9 DoF Vehicle Detection CVPRW 2020 [dataset, Daimler, mono3D]
• NYC3DCars: A Dataset of 3D Vehicles in Geographic Context ICCV 2013
• Towards Fully Autonomous Driving: Systems and Algorithms IV 2011
• Center3D: Center-based Monocular 3D Object Detection with Joint Depth Understanding [Notes] [mono3D, LID+DepJoint]
• ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection AAAI 2020 oral [mono3D]
• CenterFusion: Center-based Radar and Camera Fusion for 3D Object Detection [Notes] WACV 2021 [early fusion, camera, radar]
• 3D-LaneNet+: Anchor Free Lane Detection using a Semi-Local Representation [Notes] NeurIPS 2020 workshop [GM Israel, 3D LLD]
• LSTR: End-to-end Lane Shape Prediction with Transformers [Notes] WACV 2021 [LLD, transformers]
• PIXOR: Real-time 3D Object Detection from Point Clouds [Notes] CVPR 2018 (birds eye view)
• HDNET/PIXOR++: Exploiting HD Maps for 3D Object Detection [Notes] CoRL 2018
• CPNDet: Corner Proposal Network for Anchor-free, Two-stage Object Detection ECCV 2020 [anchor free, two stage]
• MVF: End-to-End Multi-View Fusion for 3D Object Detection in LiDAR Point Clouds [Notes] CoRL 2019 [Waymo, VoxelNet 1st author]
• Pillar-based Object Detection for Autonomous Driving [Notes] ECCV 2020
• Training-Time-Friendly Network for Real-Time Object Detection AAAI 2020 [anchor-free, fast training]
• Autonomous Driving with Deep Learning: A Survey of State-of-Art Technologies [Review of autonomous stack, Yu Huang]
• Dense Monocular Depth Estimation in Complex Dynamic Scenes CVPR 2016
• Probabilistic Future Prediction for Video Scene Understanding
• AB3D: A Baseline for 3D Multi-Object Tracking IROS 2020 [3D MOT]
• Spatial-Temporal Relation Networks for Multi-Object Tracking ICCV 2019 [MOT, feature location over time]
• Beyond Pixels: Leveraging Geometry and Shape Cues for Online Multi-Object Tracking ICRA 2018 [MOT, IIT, 3D shape]
• ST-3D: Joint Spatial-Temporal Optimization for Stereo 3D Object Tracking CVPR 2020 [Peilinag LI, author of VINS and S3DOT]
• Augment Your Batch: Improving Generalization Through Instance Repetition CVPR 2020
• RetinaTrack: Online Single Stage Joint Detection and Tracking CVPR 2020 [MOT]
• Object as Hotspots: An Anchor-Free 3D Object Detection Approach via Firing of Hotspots
• Gradient Centralization: A New Optimization Technique for Deep Neural Networks ECCV 2020 oral
• Depth Completion via Deep Basis Fitting WACV 2020
• BTS: From Big to Small: Multi-Scale Local Planar Guidance for Monocular Depth Estimation [monodepth, supervised]
• The Edge of Depth: Explicit Constraints between Segmentation and Depth CVPR 2020 [monodepth, Xiaoming Liu]
• On the Continuity of Rotation Representations in Neural Networks CVPR 2019 [rotational representation]
• VDO-SLAM: A Visual Dynamic Object-aware SLAM System IJRR 2020
• Dynamic SLAM: The Need For Speed
• Pseudo RGB-D for Self-Improving Monocular SLAM and Depth Prediction ECCV 2020
• Traffic Light Mapping and Detection [Notes] ICRA 2011 [traffic light, Google, Chris Urmson]
• Traffic light recognition exploiting map and localization at every stage [Notes] Expert Systems 2017 [traffic light, 鲜于明镐，徐在圭，郑浩奇]
• Traffic Light Recognition Using Deep Learning and Prior Maps for Autonomous Cars [Notes] IJCNN 2019 [traffic light, Espirito Santo Brazil]

2020-10 (14)

• TSM: Temporal Shift Module for Efficient Video Understanding [Notes] ICCV 2019 [Song Han, video, object detection]
• WOD: Waymo Dataset: Scalability in Perception for Autonomous Driving: Waymo Open Dataset [Notes] CVPR 2020
• Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection [Notes] NeurIPS 2020 [classification as regression]
• A Ranking-based, Balanced Loss Function Unifying Classification and Localisation in Object Detection NeurIPS 2020 spotlight
• Rethinking the Value of Labels for Improving Class-Imbalanced Learning NeurIPS 2020
• RepLoss: Repulsion Loss: Detecting Pedestrians in a Crowd [Notes] CVPR 2018 [crowd detection, Megvii]
• Adaptive NMS: Refining Pedestrian Detection in a Crowd [Notes] CVPR 2019 oral [crowd detection, NMS]
• AggLoss: Occlusion-aware R-CNN: Detecting Pedestrians in a Crowd [Notes] ECCV 2018 [crowd detection]
• CrowdDet: Detection in Crowded Scenes: One Proposal, Multiple Predictions [Notes] CVPR 2020 oral [crowd detection, Megvii, Earth mover's distance]
• R2-NMS: NMS by Representative Region: Towards Crowded Pedestrian Detection by Proposal Pairing [Notes] CVPR 2020
• Double Anchor R-CNN for Human Detection in a Crowd [Notes] [head-body bundle]
• Review: AP vs MR
• SKU110K: Precise Detection in Densely Packed Scenes [Notes] CVPR 2019 [crowd detection, no occlusion]
• GossipNet: Learning non-maximum suppression CVPR 2017
• TLL: Small-scale Pedestrian Detection Based on Somatic Topology Localization and Temporal Feature Aggregation ECCV 2018
• Learning Monocular 3D Vehicle Detection without 3D Bounding Box Labels GCPR 2020 [mono3D, Daniel Cremers, TUM]
• CubifAE-3D: Monocular Camera Space Cubification on Autonomous Vehicles for Auto-Encoder based 3D Object Detection [Notes] [mono3D, depth AE pretraining]
• Deformable DETR: Deformable Transformers for End-to-End Object Detection [Notes] ICLR 2021 [Jifeng Dai, DETR]
• ViT: An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale [Notes] ICLR 2021
• BYOL: Bootstrap your own latent: A new approach to self-supervised Learning [self-supervised]

2020-09 (15)

• SDFLabel: Autolabeling 3D Objects With Differentiable Rendering of SDF Shape Priors [Notes] CVPR 2020 oral [TRI, differentiable rendering]
• DensePose: Dense Human Pose Estimation In The Wild [Notes] CVPR 2018 oral [FAIR]
• NOCS: Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation CVPR 2019
• monoDR: Monocular Differentiable Rendering for Self-Supervised 3D Object Detection [Notes] ECCV 2020 [TRI, mono3D]
• Lift, Splat, Shoot: Encoding Images From Arbitrary Camera Rigs by Implicitly Unprojecting to 3D [Notes] ECCV 2020 [BEV-Net, Utoronto, Sanja Fidler]
• Implicit Latent Variable Model for Scene-Consistent Motion Forecasting ECCV 2020 [Uber ATG, Rachel Urtasun]
• FISHING Net: Future Inference of Semantic Heatmaps In Grids [Notes] CVPRW 2020 [BEV-Net, Mapping, Zoox]
• VPN: Cross-view Semantic Segmentation for Sensing Surroundings [Notes] RAL 2020 [Bolei Zhou, BEV-Net]
• VED: Monocular Semantic Occupancy Grid Mapping with Convolutional Variational Encoder-Decoder Networks [Notes] ICRA 2019 [BEV-Net]
• Cam2BEV: A Sim2Real Deep Learning Approach for the Transformation of Images from Multiple Vehicle-Mounted Cameras to a Semantically Segmented Image in Bird's Eye View [Notes] ITSC 2020 [BEV-Net]
• Learning to Look around Objects for Top-View Representations of Outdoor Scenes [Notes] ECCV 2018 [BEV-Net, UCSD, Manmohan Chandraker]
• A Parametric Top-View Representation of Complex Road Scenes CVPR 2019 [BEV-Net, UCSD, Manmohan Chandraker]
• FTM: Understanding Road Layout from Videos as a Whole CVPR 2020 [BEV-Net, UCSD, Manmohan Chandraker]
• KM3D-Net: Monocular 3D Detection with Geometric Constraints Embedding and Semi-supervised Training [Notes] RAL 2021 [RTM3D, Peixuan Li]
• InstanceMotSeg: Real-time Instance Motion Segmentation for Autonomous Driving [Notes] IROS 2020 [motion segmentation]
• MPV-Nets: Monocular Plan View Networks for Autonomous Driving [Notes] IROS 2019 [BEV-Net]
• Class-Balanced Loss Based on Effective Number of Samples [Notes] CVPR 2019 [Focal loss authors]
• Geometric Pretraining for Monocular Depth Estimation [Notes] ICRA 2020
• Robust Traffic Light and Arrow Detection Using Digital Map with Spatial Prior Information for Automated Driving [Notes] Sensors 2020 [traffic light, 金沢]

2020-08 (26)

• Feature-metric Loss for Self-supervised Learning of Depth and Egomotion [Notes] ECCV 2020 [feature-metric, local minima, monodepth]
• Depth-VO-Feat: Unsupervised Learning of Monocular Depth Estimation and Visual Odometry with Deep Feature Reconstruction CVPR 2018 [feature-metric, monodepth]
• MonoResMatch: Learning monocular depth estimation infusing traditional stereo knowledge [Notes] CVPR 2019 [monodepth, local minima, cheap stereo GT]
• SGDepth: Self-Supervised Monocular Depth Estimation: Solving the Dynamic Object Problem by Semantic Guidance [Notes] ECCV 2020 [Moving objects]
• Every Pixel Counts: Unsupervised Geometry Learning with Holistic 3D Motion Understanding ECCV 2018 [dynamic objects, rigid and dynamic motion]
• Every Pixel Counts ++: Joint Learning of Geometry and Motion with 3D Holistic Understanding TPAMI 2018
• CC: Competitive Collaboration: Joint Unsupervised Learning of Depth, Camera Motion, Optical Flow and Motion Segmentation [Notes] CVPR 2019
• ObjMotionNet: Self-supervised Object Motion and Depth Estimation from Video [Notes] CVPRW 2020 [object motion prediction, velocity prediction]
• Instance-wise Depth and Motion Learning from Monocular Videos
• Semantics-Driven Unsupervised Learning for Monocular Depth and Ego-Motion Estimation
• Self-Supervised Joint Learning Framework of Depth Estimation via Implicit Cues
• DF-Net: Unsupervised Joint Learning of Depth and Flow using Cross-Task Consistency ECCV 2018
• LineNet: a Zoomable CNN for Crowdsourced High Definition Maps Modeling in Urban Environments [mapping]
• Road-SLAM: Road Marking based SLAM with Lane-level Accuracy [Notes] [HD mapping]
• AVP-SLAM: Semantic Visual Mapping and Localization for Autonomous Vehicles in the Parking Lot [Notes] IROS 2020 [Huawei, HD mapping, Tong Qin, VINS author, autonomous valet parking]
• AVP-SLAM-Late-Fusion: Mapping and Localization using Semantic Road Marking with Centimeter-level Accuracy in Indoor Parking Lots [Notes] ITSC 2019
• Lane markings-based relocalization on highway ITSC 2019
• DeepRoadMapper: Extracting Road Topology from Aerial Images [Notes] ICCV 2017 [Uber ATG, NOT HD maps]
• RoadTracer: Automatic Extraction of Road Networks from Aerial Images CVPR 2018 [NOT HD maps]
• PolyMapper: Topological Map Extraction From Overhead Images [Notes] ICCV 2019 [mapping, polygon, NOT HD maps]
• HRAN: Hierarchical Recurrent Attention Networks for Structured Online Maps [Notes] CVPR 2018 [HD mapping, highway, polyline loss, Chamfer distance]
• Deep Structured Crosswalk: End-to-End Deep Structured Models for Drawing Crosswalks [Notes] ECCV 2018
• DeepBoundaryExtractor: Convolutional Recurrent Network for Road Boundary Extraction [Notes] CVPR 2019 [HD mapping, boundary, polyline loss]
• DAGMapper: Learning to Map by Discovering Lane Topology [Notes] ICCV 2019 [HD mapping, highway, forks and merges, polyline loss]
• Sparse-HD-Maps: Exploiting Sparse Semantic HD Maps for Self-Driving Vehicle Localization [Notes] IROS 2019 oral [Uber ATG, metadata, mapping, localization]
• Aerial LaneNet: Lane Marking Semantic Segmentation in Aerial Imagery using Wavelet-Enhanced Cost-sensitive Symmetric Fully Convolutional Neural Networks IEEE TGRS 2018
• Monocular Localization with Vector HD Map (MLVHM): A Low-Cost Method for Commercial IVs Sensors 2020 [Tsinghua, 3D HD maps]
• PatchNet: Rethinking Pseudo-LiDAR Representation [Notes] ECCV 2020 [SenseTime, Wanli Ouyang]
• D4LCN: Learning Depth-Guided Convolutions for Monocular 3D Object Detection [Notes] CVPR 2020 [mono3D]
• MfS: Learning Stereo from Single Images [Notes] ECCV 2020 [mono for stereo, learn stereo matching with mono]
• BorderDet: Border Feature for Dense Object Detection ECCV 2020 oral [Megvii]
• Scale-Aware Trident Networks for Object Detection ICCV 2019 [different heads for different scales]
• Learning Depth from Monocular Videos using Direct Methods
• Vid2Depth: Unsupervised Learning of Depth and Ego-Motion from Monocular Video Using 3D Geometric Constraints CVPR 2018 [Google]
• Atlas: End-to-End 3D Scene Reconstruction from Posed Images ECCV 2020
• NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections
• Supervising the new with the old: learning SFM from SFM [Notes] ECCV 2018
• Neural RGB->D Sensing: Depth and Uncertainty from a Video Camera CVPR 2019 [multi-frame monodepth]
• Don't Forget The Past: Recurrent Depth Estimation from Monocular Video [multi-frame monodepth, RNN]
• Recurrent Neural Network for (Un-)supervised Learning of Monocular VideoVisual Odometry and Depth [multi-frame monodepth, RNN]
• Exploiting temporal consistency for real-time video depth estimation ICCV 2019 [multi-frame monodepth, RNN, indoor]
• SfM-Net: Learning of Structure and Motion from Video [dynamic object, SfM]
• MB-Net: MergeBoxes for Real-Time 3D Vehicles Detection [Notes] IV 2018 [mono3D: Daimler]
• BS3D: Beyond Bounding Boxes: Using Bounding Shapes for Real-Time 3D Vehicle Detection from Monocular RGB Images [Notes] IV 2019 [mono3D, Daimler]
• 3D-GCK: Single-Shot 3D Detection of Vehicles from Monocular RGB Images via Geometrically Constrained Keypoints in Real-Time [Notes] IV 2020 [[mono3D, Daimler]
• UR3D: Distance-Normalized Unified Representation for Monocular 3D Object Detection [Notes] ECCV 2020 [mono3D]
• DA-3Det: Monocular 3D Object Detection via Feature Domain Adaptation [Notes] ECCV 2020 [mono3D]
• RAR-Net: Reinforced Axial Refinement Network for Monocular 3D Object Detection [Notes] ECCV 2020 [mono3D]

2020-07 (25)

• CenterTrack: Tracking Objects as Points [Notes] ECCV 2020 spotlight [camera based 3D MOD, MOT SOTA, CenterNet, video based object detection]
• CenterPoint: Center-based 3D Object Detection and Tracking [Notes] CVPR 2021 [lidar based 3D MOD, CenterNet]
• Tracktor: Tracking without bells and whistles [Notes] ICCV 2019 [Tracktor/Tracktor++, Laura Leal-Taixe@TUM]
• FairMOT: A Simple Baseline for Multi-Object Tracking [Notes]
• DeepMOT: A Differentiable Framework for Training Multiple Object Trackers [Notes] CVPR 2020 [trainable Hungarian, Laura Leal-Taixe@TUM]
• MPNTracker: Learning a Neural Solver for Multiple Object Tracking CVPR 2020 oral [trainable Hungarian, Laura Leal-Taixe@TUM]
• nuScenes: A multimodal dataset for autonomous driving [Notes] CVPR 2020 [dataset, point cloud, radar]
• CBGS: Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection [Notes] CVPRW 2019 [Megvii, lidar, WAD challenge winner]
• AFDet: Anchor Free One Stage 3D Object Detection and Competition solution [Notes] CVPRW 2020 [Horizon robotics, lidar, winning for Waymo challenge]
• Review of MOT and SOT [Notes]
• CrowdHuman: A Benchmark for Detecting Human in a Crowd [Notes] [megvii, pedestrian, dataset]
• WiderPerson: A Diverse Dataset for Dense Pedestrian Detection in the Wild [Notes] TMM 2019 [dataset, pedestrian]
• Tsinghua-Daimler Cyclists: A New Benchmark for Vison-Based Cyclist Detection [Notes] IV 2016 [dataset, cyclist Detection]
• Specialized Cyclist Detection Dataset: Challenging Real-World Computer Vision Dataset for Cyclist Detection Using a Monocular RGB Camera [Notes] IV 2019 [Extention to KITTI]
• PointTrack: Segment as Points for Efficient Online Multi-Object Tracking and Segmentation [Notes] ECCV 2020 oral [MOTS]
• PointTrack++ for Effective Online Multi-Object Tracking and Segmentation [Notes] CVPR 2020 workshop [CVPR2020 MOTS Challenge Winner. PointTrack++ ranks first on KITTI MOTS]
• SpatialEmbedding: Instance Segmentation by Jointly Optimizing Spatial Embeddings and Clustering Bandwidth [Notes] ICCV 2019 [one-stage, instance segmentation]
• BA-Net: Dense Bundle Adjustment Networks [Notes] ICLR 2019 [Bundle adjustment, multi-frame monodepth, feature-metric]
• DeepSFM: Structure From Motion Via Deep Bundle Adjustment ECCV 2020 oral [multi-frame monodepth, indoor scene]
• CVD: Consistent Video Depth Estimation [Notes] SIGGRAPH 2020 [multi-frame monodepth, online finetune]
• DeepV2D: Video to Depth with Differentiable Structure from Motion [Notes] ICLR 2020 [multi-frame monodepth, Jia Deng]
• GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose [Notes] CVPR 2018 [residual optical flow, monodepth, rigid and dynamic motion]
• GLNet: Self-supervised Learning with Geometric Constraints in Monocular Video: Connecting Flow, Depth, and Camera [Notes] ICCV 2019 [online finetune, rigid and dynamic motion]
• Depth Hints: Self-Supervised Monocular Depth Hints [Notes] ICCV 2019 [monodepth, local minima, cheap stereo GT]
• MonoUncertainty: On the uncertainty of self-supervised monocular depth estimation [Notes] CVPR 2020 [depth uncertainty]
• Self-Supervised Learning of Depth and Ego-motion with Differentiable Bundle Adjustment [Notes] [Bundle adjustment, xmotors.ai, multi-frame monodepth]
• Kinematic 3D Object Detection in Monocular Video [Notes] ECCV 2020 [multi-frame mono3D, Xiaoming Liu]
• VelocityNet: Camera-based vehicle velocity estimation from monocular video [Notes] CVPR 2017 workshop [monocular velocity estimation, CVPR 2017 challenge winner]
• Vehicle Centric VelocityNet: End-to-end Learning for Inter-Vehicle Distance and Relative Velocity Estimation in ADAS with a Monocular Camera [Notes] [monocular velocity estimation, monocular distance, SOTA]

2020-06 (20)

• LeGO-LOAM: Lightweight and Ground-Optimized Lidar Odometry and Mapping on Variable Terrain [Notes] IROS 2018 [lidar, mapping]
• PIE: A Large-Scale Dataset and Models for Pedestrian Intention Estimation and Trajectory Prediction [Notes] ICCV 2019
• JAAD: Are They Going to Cross? A Benchmark Dataset and Baseline for Pedestrian Crosswalk Behavior ICCV 2017
• Pedestrian Action Anticipation using Contextual Feature Fusion in Stacked RNNs BMVC 2019
• Is the Pedestrian going to Cross? Answering by 2D Pose Estimation IV 2018
• Intention Recognition of Pedestrians and Cyclists by 2D Pose Estimation ITSC 2019 [skeleton, pedestrian, cyclist intention]
• Attentive Single-Tasking of Multiple Tasks CVPR 2019
• DETR: End-to-End Object Detection with Transformers [Notes] ECCV 2020 oral [FAIR]
• Transformer: Attention Is All You Need [Notes] NIPS 2017
• SpeedNet: Learning the Speediness in Videos [Notes] CVPR 2020 oral
• MonoPair: Monocular 3D Object Detection Using Pairwise Spatial Relationships [Notes] CVPR 2020 [Mono3D, pairwise relationship]
• SMOKE: Single-Stage Monocular 3D Object Detection via Keypoint Estimation [Notes] CVPRW 2020 [Mono3D, Zongmu]
• Vehicle Re-ID for Surround-view Camera System [Notes] CVPRW 2020 [tireline, vehicle ReID, Zongmu]
• End-to-End Lane Marker Detection via Row-wise Classification [Notes] [Qualcomm Korea, LLD as cls]
• Reliable multilane detection and classification by utilizing CNN as a regression network ECCV 2018 [LLD as reg]
• SUPER: A Novel Lane Detection System [Notes]
• Learning Lightweight Lane Detection CNNs by Self Attention Distillation ICCV 2019
• StixelNet: A Deep Convolutional Network for Obstacle Detection and Road Segmentation BMVC 2015
• StixelNetV2: Real-time category-based and general obstacle detection for autonomous driving [Notes] ICCV 2017 [DS]
• Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network [Notes] CVPR 2016 [channel-to-pixel]
• Car Pose in Context: Accurate Pose Estimation with Ground Plane Constraints [mono3D]
• Self-Mono-SF: Self-Supervised Monocular Scene Flow Estimation [Notes] CVPR 2020 oral [scene-flow, Stereo input]
• MEBOW: Monocular Estimation of Body Orientation In the Wild [Notes] CVPR 2020
• VG-NMS: Visibility Guided NMS: Efficient Boosting of Amodal Object Detection in Crowded Traffic Scenes [Notes] NeurIPS 2019 workshop [Crowded scene, NMS, Daimler]
• WYSIWYG: What You See is What You Get: Exploiting Visibility for 3D Object Detection [Notes] CVPR 2020 oral [occupancy grid]
• Real-Time Panoptic Segmentation From Dense Detections [Notes] CVPR 2020 oral [bbox + semantic segmentation = panoptic segmentation, Toyota]
• Human-Centric Efficiency Improvements in Image Annotation for Autonomous Driving [Notes] CVPRW 2020 [efficient annotation]
• SurfelGAN: Synthesizing Realistic Sensor Data for Autonomous Driving [Notes] CVPR 2020 oral [Waymo, auto data generation, surfel]
• LiDARsim: Realistic LiDAR Simulation by Leveraging the Real World [Notes] CVPR 2020 oral [Uber ATG, auto data generation, surfel]
• SuMa++: Efficient LiDAR-based Semantic SLAM IROS 2019 [semantic segmentation, lidar, SLAM]
• PON/PyrOccNet: Predicting Semantic Map Representations from Images using Pyramid Occupancy Networks [Notes] CVPR 2020 oral [BEV-Net, OFT]
• MonoLayout: Amodal scene layout from a single image [Notes] WACV 2020 [BEV-Net]
• BEV-Seg: Bird’s Eye View Semantic Segmentation Using Geometry and Semantic Point Cloud [Notes] CVPR 2020 workshop [BEV-Net, Mapping]
• A Geometric Approach to Obtain a Bird's Eye View from an Image ICCVW 2019 [mapping, geometry, Andrew Zisserman]
• FrozenDepth: Learning the Depths of Moving People by Watching Frozen People [Notes] CVPR 2019 oral
• ORB-SLAM: a Versatile and Accurate Monocular SLAM System TRO 2015
• ORB-SLAM2: an Open-Source SLAM System for Monocular, Stereo and RGB-D Cameras TRO 2016
• CubeSLAM: Monocular 3D Object SLAM [Notes] TRO 2019 [dynamic SLAM, orb slam + mono3D]
• ClusterVO: Clustering Moving Instances and Estimating Visual Odometry for Self and Surroundings [Notes] CVPR 2020 [general dynamic SLAM]
• S3DOT: Stereo Vision-based Semantic 3D Object and Ego-motion Tracking for Autonomous Driving [Notes] ECCV 2018 [Peiliang Li]
• Multi-object Monocular SLAM for Dynamic Environments [Notes] IV 2020 [monolayout authors]
• PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume [Notes] CVPR 2018 oral [Optical flow]
• LiteFlowNet: A Lightweight Convolutional Neural Network for Optical Flow Estimation CVPR 2018 [Optical flow]
• FlowNet: Learning Optical Flow With Convolutional Networks ICCV 2015 [Optical flow]
• FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks CVPR 2017 [Optical flow]
• ESPNetv2: A Light-weight, Power Efficient, and General Purpose Convolutional Neural Network CVPR 2019 [semantic segmentation, lightweight]
• Mono-SF: Multi-View Geometry Meets Single-View Depth for Monocular Scene Flow Estimation of Dynamic Traffic Scenes ICCV 2019 [depth uncertainty]

2020-05 (19)

• Egocentric Vision-based Future Vehicle Localization for Intelligent Driving Assistance Systems [Notes] [Honda] ICRA 2019
• PackNet: 3D Packing for Self-Supervised Monocular Depth Estimation [Notes] CVPR 2020 oral [Scale aware depth]
• PackNet-SG: Semantically-Guided Representation Learning for Self-Supervised Monocular Depth [Notes] ICLR 2020 [TRI, infinite-depth problem]
• TrianFlow: Towards Better Generalization: Joint Depth-Pose Learning without PoseNet [Notes] CVPR 2020 [Scale aware]
• Understanding the Limitations of CNN-based Absolute Camera Pose Regression [Notes] CVPR 2019 [Drawbacks of PoseNet, MapNet, Laura Leal-Taixe@TUM]
• To Learn or Not to Learn: Visual Localization from Essential Matrices [Notes] ICRA 2020 [SIFT + 5 pt solver >> others for VO, Laura Leal-Taixe@TUM]
• DF-VO: Visual Odometry Revisited: What Should Be Learnt? [Notes] ICRA 2020 [Depth and Flow for accurate VO]
• D3VO: Deep Depth, Deep Pose and Deep Uncertainty for Monocular Visual Odometry [Notes] CVPR 2020 oral [Daniel Cremers, TUM, depth uncertainty]
• Network Slimming: Learning Efficient Convolutional Networks through Network Slimming [Notes] ICCV 2017
• BatchNorm Pruning: Rethinking the Smaller-Norm-Less-Informative Assumption in Channel Pruning of Convolution Layers [Notes] ICLR 2018
• Direct Sparse Odometry PAMI 2018
• Train in Germany, Test in The USA: Making 3D Object Detectors Generalize [Notes] CVPR 2020
• PseudoLidarV3: End-to-End Pseudo-LiDAR for Image-Based 3D Object Detection [Notes] CVPR 2020
• ATSS: Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection [Notes] CVPR 2020 oral
• Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression AAAI 2020
• Enhancing Geometric Factors in Model Learning and Inference for Object Detection and Instance Segmentation [Journal version]
• YOLOv4: Optimal Speed and Accuracy of Object Detection [Notes]
• CBN: Cross-Iteration Batch Normalization [Notes]
• Stitcher: Feedback-driven Data Provider for Object Detection [Notes]
• SKNet: Selective Kernel Networks [Notes] CVPR 2019
• CBAM: Convolutional Block Attention Module [Notes] ECCV 2018
• ResNeSt: Split-Attention Networks [Notes]

2020-04 (14)

• ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worst [Notes] RSS 2019 [Waymo]
• IntentNet: Learning to Predict Intention from Raw Sensor Data [Notes] CoRL 2018 [Uber ATG, perception and prediction, Lidar+Map]
• RoR: Rules of the Road: Predicting Driving Behavior with a Convolutional Model of Semantic Interactions [Notes] CVPR 2019 [Zoox]
• MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Prediction [Notes] CoRL 2019 [Waymo, authors from RoR and ChauffeurNet]
• NMP: End-to-end Interpretable Neural Motion Planner [Notes] CVPR 2019 oral [Uber ATG]
• Multimodal Trajectory Predictions for Autonomous Driving using Deep Convolutional Networks [Notes] ICRA 2019 [Henggang Cui, Multimodal, Uber ATG Pittsburgh]
• Uncertainty-aware Short-term Motion Prediction of Traffic Actors for Autonomous Driving WACV 2020 [Uber ATG Pittsburgh]
• Jointly Learnable Behavior and Trajectory Planning for Self-Driving Vehicles IROS 2019 Oral [Uber ATG, behavioral planning, motion planning]
• TensorMask: A Foundation for Dense Object Segmentation [Notes] ICCV 2019 [single-stage instance seg]
• BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation [Notes] CVPR 2020 oral
• Mask Encoding for Single Shot Instance Segmentation [Notes] CVPR 2020 oral [single-stage instance seg, Chunhua Shen]
• PolarMask: Single Shot Instance Segmentation with Polar Representation [Notes] CVPR 2020 oral [single-stage instance seg]
• SOLO: Segmenting Objects by Locations [Notes] ECCV 2020 [single-stage instance seg, Chunhua Shen]
• SOLOv2: Dynamic, Faster and Stronger [Notes] [single-stage instance seg, Chunhua Shen]
• CondInst: Conditional Convolutions for Instance Segmentation [Notes] ECCV 2020 oral [single-stage instance seg, Chunhua Shen]
• CenterMask: Single Shot Instance Segmentation With Point Representation [Notes]CVPR 2020

2020-03 (15)

• VPGNet: Vanishing Point Guided Network for Lane and Road Marking Detection and Recognition [Notes] ICCV 2017
• Which Tasks Should Be Learned Together in Multi-task Learning? [Notes] [Stanford, MTL] ICML 2020
• MGDA: Multi-Task Learning as Multi-Objective Optimization NeurIPS 2018
• Taskonomy: Disentangling Task Transfer Learning [Notes] CVPR 2018
• Rethinking ImageNet Pre-training [Notes] ICCV 2019 [Kaiming He]
• UnsuperPoint: End-to-end Unsupervised Interest Point Detector and Descriptor [Notes] [superpoint]
• KP2D: Neural Outlier Rejection for Self-Supervised Keypoint Learning [Notes] ICLR 2020 (pointNet)
• KP3D: Self-Supervised 3D Keypoint Learning for Ego-motion Estimation [Notes] CoRL 2020 [Toyota, superpoint]
• NG-RANSAC: Neural-Guided RANSAC: Learning Where to Sample Model Hypotheses [Notes] ICCV 2019 [pointNet]
• Learning to Find Good Correspondences [Notes] CVPR 2018 Oral (pointNet)
• RefinedMPL: Refined Monocular PseudoLiDAR for 3D Object Detection in Autonomous Driving [Notes] [Huawei, Mono3D]
• DSP: Monocular 3D Object Detection with Decoupled Structured Polygon Estimation and Height-Guided Depth Estimation [Notes] AAAI 2020 (SenseTime, Mono3D)
• Robust Lane Detection from Continuous Driving Scenes Using Deep Neural Networks (LLD, LSTM)
• LaneNet: Towards End-to-End Lane Detection: an Instance Segmentation Approach [Notes] IV 2018 (LaneNet)
• 3D-LaneNet: End-to-End 3D Multiple Lane Detection [Notes] ICCV 2019
• Semi-Local 3D Lane Detection and Uncertainty Estimation [Notes] [GM Israel, 3D LLD]
• Gen-LaneNet: A Generalized and Scalable Approach for 3D Lane Detection [Notes] ECCV 2020 [Apollo, 3D LLD]
• Long-Term On-Board Prediction of People in Traffic Scenes under Uncertainty CVPR 2018 [Egocentric prediction]
• It’s Not All About Size: On the Role of Data Properties in Pedestrian Detection ECCV 2018 [pedestrian]

2020-02 (12)

• Associative Embedding: End-to-End Learning for Joint Detection and Grouping [Notes] NIPS 2017
• Pixels to Graphs by Associative Embedding [Notes] NIPS 2017
• Social LSTM: Human Trajectory Prediction in Crowded Spaces [Notes] CVPR 2017
• Online Video Object Detection using Association LSTM [Notes] [single stage, recurrent]
• SuperPoint: Self-Supervised Interest Point Detection and Description [Notes] CVPR 2018 (channel-to-pixel, deep SLAM, Magic Leap)
• PointRend: Image Segmentation as Rendering [Notes] CVPR 2020 Oral [Kaiming He, FAIR]
• Multigrid: A Multigrid Method for Efficiently Training Video Models [Notes] CVPR 2020 Oral [Kaiming He, FAIR]
• GhostNet: More Features from Cheap Operations [Notes] CVPR 2020
• FixRes: Fixing the train-test resolution discrepancy [Notes] NIPS 2019 [FAIR]
• MoVi-3D: Towards Generalization Across Depth for Monocular 3D Object Detection [Notes] ECCV 2020 [Virtual Cam, viewport, Mapillary/Facebook, Mono3D]
• Amodal Completion and Size Constancy in Natural Scenes [Notes] ICCV 2015 (Amodal completion)
• MoCo: Momentum Contrast for Unsupervised Visual Representation Learning [Notes] CVPR 2020 Oral [FAIR, Kaiming He]

2020-01 (19)

• Double Descent: Reconciling modern machine learning practice and the bias-variance trade-of [Notes] PNAS 2019
• Deep Double Descent: Where Bigger Models and More Data Hurt [Notes]
• Visualizing the Loss Landscape of Neural Nets NIPS 2018
• The ApolloScape Open Dataset for Autonomous Driving and its Application CVPR 2018 (dataset)
• ApolloCar3D: A Large 3D Car Instance Understanding Benchmark for Autonomous Driving [Notes] CVPR 2019
• Part-level Car Parsing and Reconstruction from a Single Street View [Notes] [Baidu]
• 6D-VNet: End-to-end 6DoF Vehicle Pose Estimation from Monocular RGB Images [Notes] CVPR 2019
• RTM3D: Real-time Monocular 3D Detection from Object Keypoints for Autonomous Driving [Notes] ECCV 2020 spotlight
• DORN: Deep Ordinal Regression Network for Monocular Depth Estimation [Notes] CVPR 2018 [monodepth, supervised]
• D&T: Detect to Track and Track to Detect [Notes] ICCV 2017 (from Feichtenhofer)
• CRF-Net: A Deep Learning-based Radar and Camera Sensor Fusion Architecture for Object Detection [Notes] SDF 2019 (radar detection)
• RVNet: Deep Sensor Fusion of Monocular Camera and Radar for Image-based Obstacle Detection in Challenging Environments [Notes] PSIVT 2019
• RRPN: Radar Region Proposal Network for Object Detection in Autonomous Vehicles [Notes] ICIP 2019
• ROLO: Spatially Supervised Recurrent Convolutional Neural Networks for Visual Object Tracking [Notes] ISCAS 2016
• Recurrent SSD: Recurrent Multi-frame Single Shot Detector for Video Object Detection [Notes] BMVC 2018 (Mitsubishi)
• Recurrent RetinaNet: A Video Object Detection Model Based on Focal Loss [Notes] ICONIP 2018 (single stage, recurrent)
• Actions as Moving Points [Notes] [not suitable for online]
• The PREVENTION dataset: a novel benchmark for PREdiction of VEhicles iNTentIONs [Notes] ITSC 2019 [dataset, cut-in]
• Semi-Automatic High-Accuracy Labelling Tool for Multi-Modal Long-Range Sensor Dataset [Notes] IV 2018
• Astyx dataset: Automotive Radar Dataset for Deep Learning Based 3D Object Detection [Notes] EuRAD 2019 (Astyx)
• Astyx camera radar: Deep Learning Based 3D Object Detection for Automotive Radar and Camera [Notes] EuRAD 2019 (Astyx)

2019-12 (12)

• How Do Neural Networks See Depth in Single Images? [Notes] ICCV 2019
• Self-supervised Sparse-to-Dense: Self-supervised Depth Completion from LiDAR and Monocular Camera ICRA 2019 (depth completion)
• DC: Depth Coefficients for Depth Completion [Notes] CVPR 2019 [Xiaoming Liu, Multimodal]
• Parse Geometry from a Line: Monocular Depth Estimation with Partial Laser Observation [Notes] ICRA 2017
• VO-Monodepth: Enhancing self-supervised monocular depth estimation with traditional visual odometry [Notes] 3DV 2019 (sparse to dense)
• Probabilistic Object Detection: Definition and Evaluation [Notes]
• The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentation [Notes] ICCV 2019
• On Calibration of Modern Neural Networks [Notes] ICML 2017 (Weinberger)
• Extreme clicking for efficient object annotation [Notes] ICCV 2017
• Radar and Camera Early Fusion for Vehicle Detection in Advanced Driver Assistance Systems [Notes] NeurIPS 2019 (radar)
• Deep Active Learning for Efficient Training of a LiDAR 3D Object Detector [Notes] IV 2019
• C3DPO: Canonical 3D Pose Networks for Non-Rigid Structure From Motion [Notes] ICCV 2019
• YOLACT: Real-time Instance Segmentation [Notes] ICCV 2019 [single-stage instance seg]
• YOLACT++: Better Real-time Instance Segmentation [single-stage instance seg]

2019-11 (20)

• Review of Image and Feature Descriptors
• Vehicle Detection With Automotive Radar Using Deep Learning on Range-Azimuth-Doppler Tensors [Notes] ICCV 2019
• GPP: Ground Plane Polling for 6DoF Pose Estimation of Objects on the Road [Notes] IV 2020 [UCSD, Trevidi, mono 3DOD]
• MVRA: Multi-View Reprojection Architecture for Orientation Estimation [Notes] ICCV 2019
• YOLOv3: An Incremental Improvement
• Gaussian YOLOv3: An Accurate and Fast Object Detector Using Localization Uncertainty for Autonomous Driving [Notes] ICCV 2019 (Detection with Uncertainty)
• Bayesian YOLOv3: Uncertainty Estimation in One-Stage Object Detection [Notes] [DriveU]
• Towards Safe Autonomous Driving: Capture Uncertainty in the Deep Neural Network For Lidar 3D Vehicle Detection [Notes] ITSC 2018 (DriveU)
• Leveraging Heteroscedastic Aleatoric Uncertainties for Robust Real-Time LiDAR 3D Object Detection [Notes] IV 2019 (DriveU)
• Can We Trust You? On Calibration of a Probabilistic Object Detector for Autonomous Driving [Notes] IROS 2019 (DriveU)
• LaserNet: An Efficient Probabilistic 3D Object Detector for Autonomous Driving [Notes] CVPR 2019 (uncertainty)
• LaserNet KL: Learning an Uncertainty-Aware Object Detector for Autonomous Driving [Notes] [LaserNet with KL divergence]
• IoUNet: Acquisition of Localization Confidence for Accurate Object Detection [Notes] ECCV 2018
• gIoU: Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression [Notes] CVPR 2019
• The Lovász-Softmax loss: A tractable surrogate for the optimization of the intersection-over-union measure in neural networks CVPR 2018 [IoU as loss]
• KL Loss: Bounding Box Regression with Uncertainty for Accurate Object Detection [Notes] CVPR 2019
• CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth [Notes] CVPR 2019
• BayesOD: A Bayesian Approach for Uncertainty Estimation in Deep Object Detectors [Notes]
• TW-SMNet: Deep Multitask Learning of Tele-Wide Stereo Matching [Notes] ICIP 2019
• Accurate Uncertainties for Deep Learning Using Calibrated Regression [Notes] ICML 2018
• Calibrating Uncertainties in Object Localization Task [Notes] NIPS 2018
• SMWA: On the Over-Smoothing Problem of CNN Based Disparity Estimation [Notes] ICCV 2019 [Multimodal, depth estimation]
• Sparse-to-Dense: Depth Prediction from Sparse Depth Samples and a Single Image [Notes] ICRA 2018 (depth completion)

2019-10 (18)

• Review of monocular object detection
• Review of 2D 3D contraints in Mono 3DOD
• MonoGRNet 2: Monocular 3D Object Detection via Geometric Reasoning on Keypoints [Notes] [estimates depth from keypoints]
• Deep MANTA: A Coarse-to-fine Many-Task Network for joint 2D and 3D vehicle analysis from monocular image [Notes] CVPR 2017
• SS3D: Monocular 3D Object Detection and Box Fitting Trained End-to-End Using Intersection-over-Union Loss [Notes] [rergess distance from images, centernet like]
• GS3D: An Efficient 3D Object Detection Framework for Autonomous Driving [Notes] CVPR 2019
• M3D-RPN: Monocular 3D Region Proposal Network for Object Detection [Notes] ICCV 2019 oral [3D anchors, cyclists, Xiaoming Liu]
• TLNet: Triangulation Learning Network: from Monocular to Stereo 3D Object Detection [Notes] CVPR 2019
• A Survey on 3D Object Detection Methods for Autonomous Driving Applications [Notes] TITS 2019 [Review]
• BEV-IPM: Deep Learning based Vehicle Position and Orientation Estimation via Inverse Perspective Mapping Image [Notes] IV 2019
• ForeSeE: Task-Aware Monocular Depth Estimation for 3D Object Detection [Notes] AAAI 2020 oral [successor to pseudo-lidar, mono 3DOD SOTA]
• Obj-dist: Learning Object-specific Distance from a Monocular Image [Notes] ICCV 2019 (xmotors.ai + NYU) [monocular distance]
• DisNet: A novel method for distance estimation from monocular camera [Notes] IROS 2018 [monocular distance]
• BirdGAN: Learning 2D to 3D Lifting for Object Detection in 3D for Autonomous Vehicles [Notes] IROS 2019
• Shift R-CNN: Deep Monocular 3D Object Detection with Closed-Form Geometric Constraints [Notes] ICIP 2019
• 3D-RCNN: Instance-level 3D Object Reconstruction via Render-and-Compare [Notes] CVPR 2018
• Deep Optics for Monocular Depth Estimation and 3D Object Detection [Notes] ICCV 2019
• MonoLoco: Monocular 3D Pedestrian Localization and Uncertainty Estimation [Notes] ICCV 2019
• Joint Monocular 3D Vehicle Detection and Tracking [Notes] ICCV 2019 (Berkeley DeepDrive)
• CasGeo: 3D Bounding Box Estimation for Autonomous Vehicles by Cascaded Geometric Constraints and Depurated 2D Detections Using 3D Results [Notes]

2019-09 (17)

• Slimmable Neural Networks [Notes] ICLR 2019
• Universally Slimmable Networks and Improved Training Techniques [Notes] ICCV 2019
• AutoSlim: Towards One-Shot Architecture Search for Channel Numbers
• Once for All: Train One Network and Specialize it for Efficient Deployment
• DOTA: A Large-scale Dataset for Object Detection in Aerial Images [Notes] CVPR 2018 (rotated bbox)
• RoiTransformer: Learning RoI Transformer for Oriented Object Detection in Aerial Images [Notes] CVPR 2019 (rotated bbox)
• RRPN: Arbitrary-Oriented Scene Text Detection via Rotation Proposals TMM 2018
• R2CNN: Rotational Region CNN for Orientation Robust Scene Text Detection (rotated bbox)
• TI white paper: Webinar: mmWave Radar for Automotive and Industrial applications [Notes] [TI, radar]
• Federated Learning: Strategies for Improving Communication Efficiency [Notes] NIPS 2016
• sort: Simple Online and Realtime Tracking [Notes] ICIP 2016
• deep-sort: Simple Online and Realtime Tracking with a Deep Association Metric [Notes]
• MT-CNN: Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks [Notes] SPL 2016 (real time, facial landmark)
• RetinaFace: Single-stage Dense Face Localisation in the Wild [Notes] CVPR 2020 [joint object and landmark detection]
• SC-SfM-Learner: Unsupervised Scale-consistent Depth and Ego-motion Learning from Monocular Video [Notes] NIPS 2019
• SiamMask: Fast Online Object Tracking and Segmentation: A Unifying Approach CVPR 2019 (tracking, segmentation, label propagation)
• Review of Kálmán Filter (from Tim Babb, Pixar Animation) [Notes]
• R-FCN: Object Detection via Region-based Fully Convolutional Networks [Notes] NIPS 2016
• Guided backprop: Striving for Simplicity: The All Convolutional Net [Notes] ICLR 2015
• Occlusion-Net: 2D/3D Occluded Keypoint Localization Using Graph Networks [Notes] CVPR 2019
• Boxy Vehicle Detection in Large Images [Notes] ICCV 2019
• FQNet: Deep Fitting Degree Scoring Network for Monocular 3D Object Detection [Notes] CVPR 2019 [Mono 3DOD, Jiwen Lu]

2019-08 (18)

• Mono3D: Monocular 3D Object Detection for Autonomous Driving [Notes] CVPR2016
• MonoDIS: Disentangling Monocular 3D Object Detection [Notes] ICCV 2019
• Pseudo lidar-e2e: Monocular 3D Object Detection with Pseudo-LiDAR Point Cloud [Notes] ICCV 2019 (pseudo-lidar with 2d and 3d consistency loss, better than PL and worse than PL++, SOTA for pure mono3D)
• MonoGRNet: A Geometric Reasoning Network for Monocular 3D Object Localization [Notes] AAAI 2019 (SOTA of Mono3DOD, MLF < MonoGRNet < Pseudo-lidar)
• MLF: Multi-Level Fusion based 3D Object Detection from Monocular Images [Notes] CVPR 2018 (precursor to pseudo-lidar)
• [ROI-10D: Monocular Lifting of 2D Detection to 6