3d-detector papers within year 2021, including CVPR, ICRA, ICLR, AAAI, ICCV, ACM MM, arXiv, etc.
| Pub. | Org. | Title | Remark |
|---|---|---|---|
| ICCV2021 | CUHK, HUAWEI Noah | Pyramid R-CNN:Towards Better Performance and Adaptability for 3D Object Detection | det:kitti,Waymo |
| ICCV2021 | QCraft | Exploring Simple 3D Multi-Object Tracking for Autonomous Driving | tracking |
| ICCV2021(oral) | Tsinghua | PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers | point completion |
| ICCV2021 | CUHK, HUAWEI Noah | VoTr: Voxel Transformer for 3D Object Detection | det:kitti,Waymo |
| ICCV2021 | USC, Waymo | SPG: Unsupervised Domain Adaptation for 3D Object Detection via Semantic Point Generation | det:kitti,Waymo |
| ICCV2021 | Zhejiang University, DAMO Academy | Improving 3D Object Detection with Channel-wise Transformer | det:kitti, Waymo |
| ICCV2021 | FAIR | An End-to-End Transformer Model for 3D Object Detection | det:ScanNet |
| CVPR2021 | Southeast University, NUST | SIENet: Spatial Information Enhancement Network for 3D Object Detection from Point Cloud | det:kitti |
| CVPR2021 | Yonsei University | HVPR: Hybrid Voxel-Point Representation for Single-stage 3D Object Detection | det:kitti |
| CVPR2021 | CUHK | SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud) | det:kitti |
| CVPR2021 | Waymo, Google brain | To the Point: Efficient 3D Object Detection in the Range Image with Graph Convolution Kernels | det:Waymo |
| CVPR2021 | TuSimple | LiDAR R-CNN: An Efficient and Universal 3D Object Detector | det:kitti,Waymo |
| CVPR2021 | beihang | Back-tracing Representative Points for Voting-based 3D Object Detection in Point Clouds | indoor det |
| CVPR2021 | University of Hong Kong | ST3D: Self-training for Unsupervised Domain Adaptation on 3D Object Detection | domain adaptation:kitti,waymo,nuScenes |
| CVPR2021 | Stanford University | 3DIoUMatch: Leveraging IoU Prediction for Semi-Supervised 3D Object Detection | semi supervised det:kitti |
| CVPR2021 | CUHK | Bidirectional Projection Network for Cross Dimension Scene Understanding | 2D,3D seg |
| CVPR2021 | Tsinghua University | PV-RAFT: Point-Voxel Correlation Fields for Scene Flow Estimation of Point Clouds | estimate scene flow from point clouds |
| CVPR2021 | HKUST | FFB6D：A Full Flow Bidirectional Fusion Network for 6D Pose Estimation | RGBD:fuse RGB & D |
| CVPR2021 | CUHK | Point Cloud Upsampling via Disentangled Refinement | generative model & upsample |
| CVPR2021 | Peking University | Diffusion Probabilistic Models for 3D Point Cloud Generation | generative model & upsample(best paper finalist) |
| CVPR2021 | Stanford University | Rethinking Sampling in 3D Point Cloud GANs | generative model & upsample |
| CVPR2021WAD | Horizon Robotics | 1st Place Solution for Waymo Open Dataset Challenge | det: Waymo |
| CVPR2021 | ReLER | Point 4D Transformer Networks for Spatio-Temporal Modeling in Point Cloud Videos. | Point Cloud Videos |
| ICLR2021 | National University of Singapore | PSTNet: Point Spatio-Temporal Convolution on Point Cloud Sequences | point cloud sequences/vedio |
| ICRA2021 | BNRist, Tsinghua | FGR: Frustum-Aware Geometric Reasoning for Weakly Supervised 3D Vehicle Detection | det:kitti |
| AAAI2021 | CUHK | CIA-SSD: Confident IoU-Aware Single-Stage Object Detector From Point Cloud | det:kitti |
| AAAI2021 | USTC | Voxel R-CNN: Towards High Performance Voxel-based 3D Object Detection | det:kitti |
| arXiv 2021.3 | CUHK, Google, Waymo | 3D-MAN: 3D Multi-frame Attention Network for Object Detection | det:Waymo |
| arXiv 2021.4 | USTC,MSRA | Group-Free 3D Object Detection via Transformers | indoor det |
| arXiv 2021.4 | University of Maryland, Fudan | M3DETR: Multi-representation, Multi-scale, Mutual-relation 3D Object Detection with Transformers | det:kitti,Waymo |
| arXiv 2021.5 | Renmin University of China | Boundary-Aware 3D Object Detection from Point Clouds | det:kitti |
| arXiv 2021.5 | Zhejiang University | X-view: Non-egocentric Multi-View 3D Object Detector | det:kitti, nuScenes |
| arXiv 2021.6 | Baidu | FusionPainting: Multimodal Fusion with Adaptive Attention for 3D Object Detection | det:nuScenes |
| arXiv 2021.6 | Google, Waymo | RSN: Range Sparse Net for Efficient, Accurate LiDAR 3D Object Detection | det:Waymo |
| arXiv 2021.7 | Google, Waymo | Lidar Light Scattering Augmentation (LISA): Physics-based Simulation of AdverseWeather Conditions for 3D Object Detection | 3D augmentation |
| arXiv 2021.7 | HUKST | DV-Det: Efficient 3D Point Cloud Object Detection with Dynamic Voxelization | det:kitti |
| arXiv 2021.8 | Horizon | Real-Time Anchor-Free Single-Stage 3D Detection with IoU-Awareness | det:Waymo |
| arXiv 2021.8 | Xi'an Jiaotong University | MBDF-Net: Multi-Branch Deep Fusion Network for 3D Object Detection | det:kitti |
| arXiv 2021.8 | Baidu | AutoShape: Real-Time Shape-Aware Monocular 3D Object Detection | det:kitti |
| ACM MM21 | Zhejiang University | From Voxel to Point: IoU-guided 3D Object Detection for Point Cloud with Voxel-to-Point Decoder | det:kitti |
| ACM MM21 | Zhejiang University | Anchor-free 3D Single Stage Detector with Mask-Guided Attention for Point Cloud | det:kitti |
| Pub. | Title. | car moderate AP(test) | Remark |
|---|---|---|---|
| ICCV2021 | Pyramid R-CNN:Towards Better Performance and Adaptability for 3D Object Detection | 82.08 | - |
| ICCV2021 | VoTr: Voxel Transformer for 3D Object Detection | 82.09 | Voxel-Transformer |
| ICCV2021 | Improving 3D Object Detection with Channel-wise Transformer | 81.77 | Transformer |
| ICCV2021 | SPG: Unsupervised Domain Adaptation for 3D Object Detection via Semantic Point Generation | 82.13 | Domain adapation |
| CVPR2021 | SIENet: Spatial Information Enhancement Network for 3D Object Detection from Point Cloud | 81.71 | - |
| CVPR2021 | HVPR: Hybrid Voxel-Point Representation for Single-stage 3D Object Detection | 77.92 | 36.1Hz |
| CVPR2021 | SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud) | 82.57 | 30.56ms |
| CVPR2021 | LiDAR R-CNN: An Efficient and Universal 3D Object Detector | 74.21 | based on pointpillar(increase 1.4mAP) |
| ICRA2021 | FGR: Frustum-Aware Geometric Reasoning for Weakly Supervised 3D Vehicle Detection | 68.47 | without 3D label |
| AAAI2021 | CIA-SSD: Confident IoU-Aware Single-Stage Object Detector From Point Cloud | 80.28 | 32FPS |
| AAAI2021 | Voxel R-CNN: Towards High Performance Voxel-based 3D Object Detection | 81.62 | 25FPS |
| arXiv 2021.4 | M3DETR: Multi-representation, Multi-scale, Mutual-relation 3D Object Detection with Transformers | 81.73 | Transformers based |
| arXiv 2021.5 | Boundary-Aware 3D Object Detection from Point Clouds | 81.61 | car bev rank 1st |
| arXiv 2021.5 | X-view: Non-egocentric Multi-View 3D Object Detector | 81.35 | Non-egocentric Multi-View |
| arXiv 2021.7 | DV-Det: Efficient 3D Point Cloud Object Detection with Dynamic Voxelization | 76.74 | 75FPS |
| arXiv 2021.8 | MBDF-Net: Multi-Branch Deep Fusion Network for 3D Object Detection | 80.00 | multi-modal |
| arXiv 2021.8 | AutoShape: Real-Time Shape-Aware Monocular 3D Object Detection | 14.17 | mono |
| ACM MM21 | From Voxel to Point: IoU-guided 3D Object Detection for Point Cloud with Voxel-to-Point Decoder | 81.58 | - |
| ACM MM21 | Anchor-free 3D Single Stage Detector with Mask-Guided Attention for Point Cloud | 79.68 | - |
| Pub. | Title. | mAP(val) | NDS(val) |
|---|---|---|---|
| arXiv 2021.5 | X-view: Non-egocentric Multi-View 3D Object Detector | 51.69 | 63.76 |
| arXiv 2021.6 | FusionPainting: Multimodal Fusion with Adaptive Attention for 3D Object Detection | 66.53 | 70.68 |
| Pub. | Title. | vehicle 3D APH L2(val) | Pedestrian 3D APH L2(val) |
|---|---|---|---|
| ICCV2021 | Pyramid R-CNN:Towards Better Performance and Adaptability for 3D Object Detection | 66.68 | - |
| ICCV2021 | VoTr: Voxel Transformer for 3D Object Detection | 65.29 | - |
| ICCV2021 | Improving 3D Object Detection with Channel-wise Transformer | 69.04 | - |
| ICCV2021 | SPG: Unsupervised Domain Adaptation for 3D Object Detection via Semantic Point Generation | 65.98 | 57.68 |
| CVPR2021 | To the Point: Efficient 3D Object Detection in the Range Image with Graph Convolution Kernels | 56.7 | 61.5 |
| CVPR2021 | LiDAR R-CNN: An Efficient and Universal 3D Object Detector | 64.2 | 51.7 |
| CVPR2021WAD | 1st Place Solution for Waymo Open Dataset Challenge | 77.83 | 76.6 |
| arXiv 2021.3 | 3D-MAN: 3D Multi-frame Attention Network for Object Detection | 67.14 | 59.04 |
| arXiv 2021.4 | M3DETR: Multi-representation, Multi-scale, Mutual-relation 3D Object Detection with Transformers | 66.02 | - |
| arXiv 2021.6 | RSN: Range Sparse Net for Efficient, Accurate LiDAR 3D Object Detection | 69.5 | 69.9 |
| arXiv 2021.8 | Real-Time Anchor-Free Single-Stage 3D Detection with IoU-Awareness | 73.89 | 72.41 |