/Awesome-3D-Detectors

Paperlist of awesome 3D detection methods

List of 3D detection methods

This is a paper and code list of some awesome 3D detection methods. We mainly collect LiDAR-involved methods in autonomous driving. It is worth noticing that we include both official and unofficial codes for each paper.

paperlist-map

News

2022.2.17 Add the link for every paper. (Happy Chinese New Year!)

Paper list

Title code Pub. Input
MV3D (Multi-View 3D Object Detection Network for Autonomous Driving) CVPR2017 I+L
F-PointNet (Frustum PointNets for 3D Object Detection from RGB-D Data) code CVPR2018 I+L
VoxelNet (VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection) CVPR2018 L
PIXOR (PIXOR: Real-time 3D Object Detection from Point Clouds) code CVPR2018 L
AVOD (Joint 3D Proposal Generation and Object Detection from View Aggregation) code IROS2018 I+L
ContFusion (Deep Continuous Fusion for Multi-Sensor 3D Object Detection) ECCV2018 I+L
SECOND (SECOND: Sparsely Embedded Convolutional Detection) code Sensors 2018 L
Complex-YOLO (Complex-YOLO: Real-time 3D Object Detection on Point Clouds) code Axiv2018 L
FBFFusing Bird’s Eye View LIDAR Point Cloud and Front View Camera Image for Deep Object Detection code IV2018 I+L
RoarNet (RoarNet: A Robust 3D Object Detection based on Region Approximation Refinement) code IV2019 I+L
PVCNN (Point-Voxel CNN for Efficient 3D Deep Learning) code NIPS2019 L
MMF(Multi-Task Multi-Sensor Fusion for 3D Object Detection) code CVPR2019 I+L
PointPillars (PointPillars: Fast Encoders for Object Detection from Point Clouds) code CVPR2019 L
Point RCNN (PointRCNN: 3D Object Proposal Generation and Detection from Point Cloud) code CVPR2019 L
LaserNet (LaserNet: An Efficient Probabilistic 3D Object Detector for Autonomous Driving) CVPR2019 L
LaserNet++ (Sensor Fusion for Joint 3D Object Detection and Semantic Segmentation) CVPR2019 I+L
**Fast PointRCNN **(Fast PointRCNN) ICCV2019 L
STD (STD: Sparse-to-Dense 3D Object Detector for Point Cloud) ICCV2019 L
VoteNet (Deep Hough Voting for 3D Object Detection in Point Clouds) code ICCV2019 L
MVX-Net (MVX-Net: Multimodal VoxelNet for 3D Object Detection) code ICRA2019 I+L
Patchs (Patch Refinement - Localized 3D Object Detection) Arxiv2019 L
StarNet (StarNet: Targeted Computation for Object Detection in Point Clouds) code Arxiv2019 L
F-ConvNet (Frustum ConvNet: Sliding Frustums to Aggregate Local Point-Wise Features for Amodal 3D Object Detection) IROS2019 I+L
PI-RCNNAn Efficient Multi-sensor 3D Object Detector with Point-based Attentive Cont-conv Fusion Module AAAI2020 I+L
TANet (TANet: Robust 3D Object Detection from Point Clouds with Triple Attention) code AAAI2020 L
MVF (End-to-end multi-view fusion for 3d object detection in lidar point clouds) code ICRL2020 L
SegVoxelNet (SegVoxelNet: Exploring Semantic Context and Depth-aware Features for 3D Vehicle Detection from Point Cloud) ICRA2020 L
Voxel-FPN (Voxel-FPN: multi-scale voxel feature aggregation in 3D object detection from point clouds) Sensors 2020 L
AA3D (Adaptive and Azimuth-Aware Fusion Network of Multimodal Local Features for 3D Object Detection) Neurocomputing2020 I+L
Part A^2 (From Points to Parts: 3D Object Detection From Point Cloud With Part-Aware and Part-Aggregation Network) code TPAMI2020 L
PV-RCNN (PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection) code CVPR2020 L
3D SSD (3DSSD: Point-based 3D Single Stage Object Detector) code CVPR2020 L
Associate-3Ddet (Associate-3Ddet: Perceptual-to-Conceptual Association for 3D Point Cloud Object Detection) code CVPR2020 L
HVNet (HVNet: Hybrid Voxel Network for LiDAR Based 3D Object Detection) code CVPR2020 L
ImVoteNet (ImVoteNet: Boosting 3D Object Detection in Point Clouds with Image Votes) CVPR2020 I+L
Point GNN (Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud) CVPR2020 L
SA-SSD (Structure Aware Single-stage 3D Object Detection from Point Cloud) code CVPR2020 L
(What You See is What You Get: Exploiting Visibility for 3D Object Detection) CVPR2020 L
DOPS (DOPS: Learning to Detect 3D Objects and Predict their 3D Shapes) CVPR2020 L
3D IoU-Net (3D IoU-Net: IoU Guided 3D Object Detector for Point Clouds) Arxiv2020 L
3D CVF (3D-CVF: Generating Joint Camera and LiDAR Features Using Cross-View Spatial Feature Fusion for 3D Object Detection) ECCV2020 I+L
HotSpotNet (Object as Hotspots: An Anchor-Free 3D Object Detection Approach via Firing of Hotspots) ECCV2020 L
EPNet: (EPNet: Enhancing Point Features with Image Semantics for 3D Object Detection) code ECCV2020 I+L
WS3D (Weakly Supervised 3D Object Detection from Lidar Point Cloud) code ECCV2020 L
Pillar-OD Pillar-based Object Detection for Autonomous Driving code ECCV2020 L
SSN (SSN: Shape Signature Networks for Multi-class Object Detection from Point Clouds) Arxiv2020 L
CenterPoint (Center-based 3D Object Detection and Tracking) code Arxiv2020 L
AFDet (AFDet: Anchor Free One Stage 3D Object Detection) Waymo2020 L
LGR-Net (Local Grid Rendering Networks for 3D Object Detection in Point Clouds) arxiv2020.07 L
CenterNet3D (CenterNet3D:An Anchor free Object Detector for Autonomous Driving) code arxiv2020.07 L
RCD (Range Conditioned Dilated Convolutions for Scale Invariant 3D Object Detection) CoRL 2020 L
VS3D (Weakly Supervised 3D Object Detection from Point Clouds) code ACM MM2020 I+L
RangeRCNN (RangeRCNN: Towards Fast and Accurate 3D Object Detection with Range Image Representation) arxiv2020.09 L
MVAF-Net (Multi-View Adaptive Fusion Network for 3D Object Detection) arxiv2020.11 I+L
CADNet (Context-Aware Dynamic Feature Extraction for 3D Object Detection in Point Clouds) TITS 2021.05 L
DA-PointRCNN (A Density-Aware PointRCNN for 3D Objection Detection in Point Clouds) axiv2020.09 L
CVCNet(Every View Counts: Cross-View Consistency in 3D Object Detection with Hybrid-Cylindrical-Spherical Voxelization) NIPS2020 L
CIA-SSDCIA-SSD: Confident IoU-Aware Single-Stage Object Detector From Point Cloud code AAAI2021 L
IAAYIt's All Around You: Range-Guided Cylindrical Network for 3D Object Detection arxiv2020 L
SA-Det3D (Self-Attention Based Context-Aware 3D Object Detection) code arxiv2020 L
RangeDet(RangeDet: In Defense of Range View for LiDAR-based 3D Object Detection) ICCV2021 L
HVPR(HVPR: Hybrid Voxel-Point Representation for Single-stage 3D Object Detection) CVPR2021 L
SE-SSD(SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud) CVPR2021 L
PPC(To the Point: Efficient 3D Object Detection in the Range Image with Graph Convolution Kernels) CVPR2021 L
SIENet (SIENet: Spatial Information Enhancement Network for 3D Object Detection from Point Cloud) arxiv2021.04 L
PolarStream(PolarStream: Streaming Lidar Object Detection and Segmentation with Polar Pillars) arxiv2021.06 L
DV-Det(DV-Det: Efficient 3D Point Cloud Object Detection with Dynamic Voxelization) code arxiv2021.07 L
VPFNetVPFNet: Improving 3D Object Detection with Virtual Point based LiDAR and Stereo Data Fusion arxiv2021.11 I+L
LC-MV (Multi-View Fusion of Sensor Data for Improved Perception and Prediction in Autonomous Driving) WACV2022 I+L
BtcDet (Behind the Curtain: Learning Occluded Shapes for 3D Object Detection) code AAAI2022 L
To be continued...

Code list

  • mmdetection3d in pytorch

    Methods supported: SECOND, PointPillars, FreeAnchor, VoteNet, Part-A2, MVXNet

    Benchmark supported: KITTI, nuScenes, Lyft, ScanNet, SUNRGBD

  • OpenPCDet: An open source project for LiDAR-based 3D scene perception in Pytorch.

    Methods supported : PointPillars, SECOND, Part A^2, PV-RCNN, PointRCNN(ongoing).

    Benchmark supported: KITTI, Waymo (ongoing).

  • Det3d: A general 3D Object Detection codebase in PyTorch.

    Methods supported : PointPillars, SECOND, PIXOR.

    Benchmark supported: KITTI, nuScenes, Lyft.

  • second.pytorch: SECOND detector in Pytorch.

    Methods supported : PointPillars, SECOND.

    Benchmark supported: KITTI, nuScenes.

  • CenterPoint: "Center-based 3D Object Detection and Tracking" in Pytorch.

    Methods supported : CenterPoint-Pillar, Center-Voxel.

    Benchmark supported: nuScenesWaymo.

  • SA-SSD: "SA-SSD: Structure Aware Single-stage 3D Object Detection from Point Cloud" in pytorch

    Methods supported : SA-SSD.

    Benchmark supported: KITTI.

  • 3DSSD: "Point-based 3D Single Stage Object Detector " in Tensorflow.

    Methods supported : 3DSSD, PointRCNN, STD (ongoing).

    Benchmark supported: KITTI, nuScenes (ongoing).

  • Point-GNN: "Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud" in Tensorflow.

    Methods supported : Point-GNN.

    Benchmark supported: KITTI.

  • TANet: "TANet: Robust 3D Object Detection from Point Clouds with Triple Attention" in Pytorch.

    Methods supported : TANet (PointPillars, Second).

    Benchmark supported: KITTI.

  • Complex-YOLOv4-pytorch: " Complex-YOLO: Real-time 3D Object Detection on Point Clouds)" in pytorch.

    Methods supported : YOLO

    Benchmark supported: KITTI.

  • EPNet: "EPNet: Enhancing Point Features with Image Semantics for 3D Object Detection "

    Methods supported: EPNet

    Benchmark supported: KITTI, SUN-RGBD

  • Super Fast and Accurate 3D Detector:"Super Fast and Accurate 3D Object Detection based on 3D LiDAR Point Clouds"

    Benchmark supported: KITTI

Dataset list

(reference: https://mp.weixin.qq.com/s/3mpbulAgiwi5J66MzPNpJA from WeChat official account: "CNNer")