Awesome-Autonomous-Driving-Papers

---

This repository provides awesome research papers for autonomous driving perception.
I have tried my best to keep this repository up to date. If you do find a problem or have any suggestions, please raise this as an issue or make a pull request with information (format of the repo): Research paper title, datasets, metrics, objects, source code, publisher, and year.

This summary is categorized into:

• Datasets
• LiDAR-based 3D Object Detection
  • Single-stage detectors
  • Two-stage detectors
• Monocular Image-based 3D Object Detection
• LiDAR and RGB Images fusion
• Pseudo-LiDAR
• Training tricks

Abbreviations

• AP-2D: Average Precision for 2D detection (on RGB-image space)
• AP-3D: Average Precision for 3D detection
• AP-BEV: Average Precision for Birds Eye View
• AOS: Average Orientation Similarity (if 2D bounding box available)

Datasets

• KITTI
• Waymo Open Dataset
• nuScenes
• Lift

1. LiDAR-based 3D Object Detection

1.1 Single-stage detectors

Research Paper	Datasets	Metrics	Objects	Source Code	Publisher	Year
“HorizonLiDAR3D”: 1st Place Solution for Waymo Open Dataset Challenge - 3D Detection and Domain Adaptation	Waymo	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	---	ArXiv	2020
Structure Aware Single-stage 3D Object Detection from Point Cloud (SA-SSD)	KITTI	AP-3D AP-BEV	Cars	PyTorch	CVPR	2020
3DSSD: Point-based 3D Single Stage Object Detector	KITTI nuScenes	AP-3D AP-BEV	Cars	PyTorch	CVPR	2020
LaserNet: An Efficient Probabilistic 3D Object Detector for Autonomous Driving	KITTI ATG4D	AP-3D AP-BEV	Cars	---	CVPR	2019
PointPillars: Fast Encoders for Object Detection from Point Clouds	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch	CVPR	2019
SECOND: Sparsely Embedded Convolutional Detection	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch	Sensors	2018
Complex-YOLO: an euler-region-proposal for real-time 3d object detection on point clouds	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch	ECCV	2018
YOLO3D: End-to-end real-time 3D Oriented Object Bounding Box Detection from LiDAR Point Cloud	KITTI	mAP-3D	Cars, Pedestrians, Cyclists	PyTorch	ECCV	2018
Pixor: Real-time 3d object detection from point clouds	KITTI	AP-3D	Cars	PyTorch	CVPR	2018
VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch Tensorflow	CVPR	2018
3D Fully Convolutional Network using PointCloud data for Vehicle Detection	KITTI	AP-3D AOS	Cars	Tensorflow	IROS	2017

1.2 Two-stage detectors

Research Paper	Datasets	Metrics	Objects	Source Code	Publisher	Year
PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection	KITTI Waymo	AP-3D	Cars	PyTorch	CVPR	2020
Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud	KITTI	AP-3D AP-BEV	Cars (1 model), Pedestrians and Cyclists(1 model)	Tensorflow	CVPR	2020
3D IoU-Net: IoU Guided 3D Object Detector for Point Clouds	KITTI	AP-3D AP-BEV	Cars	---	ArXiv	2020
PointRCNN: 3D Object Proposal Generation and Detection from Point Cloud	KITTI	AP-3D	Cars, Pedestrians, Cyclists	PyTorch	CVPR	2019
Patch Refinement - Localized 3D Object Detection	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	---	ArXiv	2019
StarNet: Targeted Computation for Object Detection in Point Clouds	KITTI Waymo	AP-3D	Cars, Pedestrians, Cyclists	PyTorch	ArXiv	2019
Part-A^2 Net: 3D Part-Aware and Aggregation Neural Network for Object Detection from Point Cloud	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch	ArXiv	2019
Fast Point R-CNN	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	---	ICCV	2019
STD: Sparse-to-Dense 3D Object Detector for Point Cloud	KITTI	AP-3D	Cars, Pedestrians, Cyclists	---	ICCV	2019
Three-dimensional Backbone Network for 3D Object Detection in Traffic Scenes	KITTI	AP-3D	Cars, Pedestrians, Cyclists	PyTorch	ICCV	2019
Frustum ConvNet: Sliding Frustums to Aggregate Local Point-Wise Features for Amodal 3D Object Detection	KITTI SUN-RGBD	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch	IROS	2019
Frustum PointNets for 3D Object Detection from RGB-D Data	KITTI SUN-RGBD	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	Tensorflow	CVPR	2018

2. Monocular Image-based 3D Object Detection

Research Paper	Datasets	Metrics	Objects	Source Code	Publisher	Year
RTM3D: Real-time Monocular 3D Detection from Object Keypoints for Autonomous Driving	KITTI	AP-3D AP-BEV AOS	Cars	PyTorch	ECCV	2020
Stereo R-CNN based 3D Object Detection for Autonomous Driving	KITTI	AP-3D/BEV AP-2D	Cars	PyTorch	CVPR	2019
M3D-RPN: Monocular 3D Region Proposal Network for Object Detection	KITTI	AP-3D/BEV AP-2D	Cars, Pedestrians, Cyclists	PyTorch	ICCV	2019
Mono3D++: Monocular 3D Vehicle Detection with Two-Scale 3D Hypotheses and Task Priors	KITTI	AP-3D/BEV AP-2D	Cars, Pedestrians, Cyclists	---	ArXiv	2019
3D Bounding Box Estimation Using Deep Learning and Geometry	KITTI	AP	Cars, Cyclists	PyTorch	CVPR	2017
Deep MANTA: A Coarse-to-fine Many-Task Network for joint 2D and 3D vehicle analysis from monocular image	KITTI		Cars	---	CVPR	2017
Deep MANTA: A Coarse-to-fine Many-Task Network for joint 2D and 3D vehicle analysis from monocular image	KITTI	AP-2D/3D AOS	Cars	Link	ICRA	2017

3. LiDAR and RGB Images Fusion

Research Paper	Datasets	Metrics	Objects	Source Code	Publisher	Year
ImVoteNet: Boosting 3D Object Detection in Point Clouds with Image Votes	SUN RGB-D	AP-3D	37 object categories	PyTorch	CVPR	2020
Multi-Task Multi-Sensor Fusion for 3D Object Detection	KITTI	AP-2D AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch	CVPR	2019

4. Pseudo-LiDAR

Research Paper	Datasets	Metrics	Objects	Source Code	Publisher	Year
Pseudo-LiDAR++: Accurate Depth for 3D Object Detection in Autonomous Driving	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch	ICLR	2020
Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	PyTorch	CVPR	2019

5. Training tricks

Research Paper	Datasets	Metrics	Objects	Source Code	Publisher	Year
PPBA: Improving 3D Object Detection through Progressive Population Based Augmentation	KITTI	AP-3D AP-BEV	Cars, Pedestrians, Cyclists	---	ArXiv	2020
Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection	KITTI	AP-3D AP-BEV	10 object categories	PyTorch	ArXiv	2019
Weighted Point Cloud Augmentation for Neural Network Training Data Class-Imbalance	ScanNet Semantic3D			---	ArXiv	2019

6. Object tracking (in progress)

To do list:

• Add 3D object detection papers based on LiDAR/monocular images/fusion/pseudo-LiDAR.
• Add training tricks papers
• Add object tracking papers.
• Provide download.py script to automatically download .pdf files.

References

• The format of the README has been referred from RedditSota/state-of-the-art-result-for-machine-learning-problems