BEV

Papers

Vision-Centric BEV Perception: A Survey

• arxiv: https://arxiv.org/abs/2208.02797
• github: https://github.com/4DVLab/Vision-Centric-BEV-Perception

Multi-Camera 3D Object Detection

Lift, Splat, Shoot: Encoding Images From Arbitrary Camera Rigs by Implicitly Unprojecting to 3D

• intro: ECCV 2020
• intro: NVIDIA, Vector Institute, University of Toronto
• project page: https://nv-tlabs.github.io/lift-splat-shoot/
• arxiv: https://arxiv.org/abs/2008.05711
• github: https://github.com/nv-tlabs/lift-splat-shoot

BEVDet: High-Performance Multi-Camera 3D Object Detection in Bird-Eye-View

• intro: PhiGent Robotics
• arxiv: https://arxiv.org/abs/2112.11790

BEVDet4D: Exploit Temporal Cues in Multi-camera 3D Object Detection

• intro: PhiGent Robotics
• arxiv: https://arxiv.org/abs/2203.17054

BEVerse: Unified Perception and Prediction in Birds-Eye-View for Vision-Centric Autonomous Driving

• intro: Tsinghua University & PhiGent Robotics
• arxiv: https://arxiv.org/abs/2205.09743
• github: https://github.com/zhangyp15/BEVerse

BEVFormer: Learning Bird’s-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers

• intro: Nanjing University & Shanghai AI Laboratory & The University of Hong Kong
• arxiv: https://arxiv.org/abs/2203.17270
• github: https://github.com/zhiqi-li/BEVFormer

HFT: Lifting Perspective Representations via Hybrid Feature Transformation

• intro: Institute of Automation, Chinese Academy of Sciences & PhiGent Robotics
• arxiv: https://arxiv.org/abs/2204.05068
• github: https://github.com/JiayuZou2020/HFT

M^2BEV: Multi-Camera Joint 3D Detection and Segmentation with Unified Birds-Eye View Representation

• project page: https://xieenze.github.io/projects/m2bev/
• arxiv: https://arxiv.org/abs/2204.05088

BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird’s-Eye View Representation

• project page: https://bevfusion.mit.edu/
• arxiv: https://arxiv.org/abs/2205.13542
• github: https://github.com/mit-han-lab/bevfusion

BEVFusion: A Simple and Robust LiDAR-Camera Fusion Framework

• intro: Peking University & Alibaba Group
• arxiv: https://arxiv.org/abs/2205.13790
• github: https://github.com/ADLab-AutoDrive/BEVFusion

A Simple Baseline for BEV Perception Without LiDAR

• intro: Carnegie Mellon University & Toyota Research Institute
• project page: http://www.cs.cmu.edu/~aharley/bev/
• arxiv: https://arxiv.org/abs/2206.07959

BEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object Detection

• intro: Megvii Inc. (Face++) & Huazhong University of Science and Technology & Xi’an Jiaotong University
• arxiv: https://arxiv.org/abs/2206.10092

PolarFormer: Multi-camera 3D Object Detection with Polar Transformers

• intro: 1Fudan University & CASIA & Alibaba DAMO Academy & University of Surrey
• arxiv: https://arxiv.org/abs/2206.15398
• github: https://github.com/fudan-zvg/PolarFormer

ORA3D: Overlap Region Aware Multi-view 3D Object Detection

• intro: Korea University & KAIST & Hyundai Motor Company R&D Division
• arxiv: https://arxiv.org/abs/2207.00865

MSMDFusion: Fusing LiDAR and Camera at Multiple Scales with Multi-Depth Seeds for 3D Object Detection

• intro: Fudan University & Meituan
• arxiv: https://arxiv.org/abs/2209.03102

HD Map Construction

HDMapNet: An Online HD Map Construction and Evaluation Framework

• intro: ICRA 2022
• intro: Tsinghua University & MIT & Li Auto
• project page: https://tsinghua-mars-lab.github.io/HDMapNet/
• arxiv: https://arxiv.org/abs/2107.06307
• github: https://github.com/Tsinghua-MARS-Lab/HDMapNet

VectorMapNet: End-to-end Vectorized HD Map Learning

• intro: Tsinghua University & MIT & Li Auto
• arxiv: https://arxiv.org/abs/2206.08920

UniFormer: Unified Multi-view Fusion Transformer for Spatial-Temporal Representation in Bird’s-Eye-View

• intro: Zhejiang University & DJI & Shanghai AI Lab
• arxiv: https://arxiv.org/abs/2207.08536

MapTR: Structured Modeling and Learning for Online Vectorized HD Map Construction

• intro: University of Science & Technology, Horizon Robotics
• arxiv: https://arxiv.org/abs/2208.14437
• gihtub: https://github.com/hustvl/MapTR

Semantic Segmentation

LaRa: Latents and Rays for Multi-Camera Bird’s-Eye-View Semantic Segmentation

• intro: Valeo.ai & Inria
• arxiv: https://arxiv.org/abs/2206.13294

CoBEVT: Cooperative Bird’s Eye View Semantic Segmentation with Sparse Transformers

• intro: University of California, Los Angeles & University of Texas at Austin & University of California
• arxiv: https://arxiv.org/abs/2207.02202