ECCV 2022在自动驾驶方面有哪些工作？

作者 | 冯偲  编辑 | 汽车人

原文链接：zhuanlan.zhihu.com/p/590163934

点击下方卡片，关注“自动驾驶之心”公众号

ADAS巨卷干货，即可获取

点击进入→自动驾驶之心【全栈算法】技术交流群

后台回复【ECCV2022】获取ECCV2022所有自动驾驶方向论文！

自动驾驶数据集

ECCV2022|数据集--Corner Case |CODA: A Real-World Road Corner Case Dataset for Object Detection in Autonomous Driving|

ECCV2022|数据集--多目标跟踪 | Large Scale Real-World Multi-Person Tracking|

End-End自动驾驶

ECCV2022|端到端自动驾驶-BEV架构|ST-P3: End-to-end Vision-based Autonomous Driving via Spatial-Temporal Feature Learning|

ECCV2022|端到端感知预测--无监督 |Motion Inspired Unsupervised Perception and Prediction in Autonomous Driving|

BEV感知算法

ECCV2022|BEV--预测 |StretchBEV: Stretching Future Instance Prediction Spatially and Temporally|

ECCV2022|BEV感知|BEVFormer: Learning Bird’s-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers|

ECCV2022|BEV感知-目标检测|Learning Ego 3D Representation as Ray Tracing|

感知数据处理

ECCV2022|数据集处理-长尾+不均衡|On Multi-Domain Long-Tailed Recognition, Imbalanced Domain Generalization and Beyond|

ECCV2022|数据处理--点云压缩 |Point Cloud Compression with Range Image-based Entropy Model for Autonomous Driving|

ECCV2022|数据处理--闭环训练 |Rethinking Closed-loop Training for Autonomous Driving|

ECCV2022|数据增强--点云 | PseudoAugment: Learning to Use Unlabeled Data for Data Augmentation in Point Clouds|

Corner Case(Anomaly Detection)解决方案

ECCV2022|异常检测|DenseHybrid: Hybrid Anomaly Detection for Dense Open-set Recognition|

ECCV2022|异常检测/分割|Pixel-wise Energy-biased Abstention Learning for Anomaly Segmentation on Complex Urban Driving Scenes|

障碍物轨迹预测

ECCV2022|轨迹预测 |View Vertically: A Hierarchical Network for Trajectory Prediction via Fourier Spectrums|

ECCV2022|轨迹预测 |Aware of the History: Trajectory Forecasting with the Local Behavior Data|

ECCV2022|轨迹预测 |D2-TPred: Discontinuous Dependency for Trajectory Prediction under Traffic Lights|

ECCV2022|轨迹预测 |AdvDO: Realistic Adversarial Attacks for Trajectory Prediction|

ECCV2022|车辆意图预测--变道|DLane Change Classification and Prediction with Action Recognition Networks

ECCV2022|行为决策|InAction: Interpretable Action Decision Making for Autonomous Driving|

ECCV2022|轨迹预测|Action-based Contrastive Learning for Trajectory Prediction|

ECCV2022|障碍物轨迹预测|View Vertically: A Hierarchical Network for Trajectory Prediction via Fourier Spectrums|

建图与定位

ECCV2022|相机姿态估计 | SC-wLS: Towards Interpretable Feed-forward Camera Re-localization|

ECCV2022|视觉定位 |IIs Geometry Enough for Matching in Visual Localization?|

多模态数据融合算法

ECCV2022|多模态融合--目标检测 |Lidar Point Cloud Guided Monocular 3D Object Detection|

ECCV2022|多模态数据--目标检测 |CHomogeneous Multi-modal Feature Fusion and Interaction for 3D Object Detection|

ECCV2022|多模态数据融合检测 |CramNet: Camera-Radar Fusion with Ray-Constrained Cross-Attention for Robust 3D Object Detection|

多任务学习网络

ECCV2022|多任务网络(深度估计+视觉里程计+BEV)|JPerceiver: Joint Perception Network for Depth, Pose and Layout Estimation in Driving Scenes|

2d/3d目标检测算法

ECCV2022|目标检测--毫米波雷达 |Radatron: Accurate Detection Using Multi-Resolution Cascaded MIMO Radar|

ECCV2022|长尾数据--检测 |Improving the Intra-class Long-tail in 3D Detection via Rare Example Mining|

ECCV2022|3D目标检测--点云 |PillarNet: Real-Time and High-Performance Pillar-based 3D Object Detection|

ECCV2022|3D目标检测--点云 |INT: Towards Infinite-frames 3D Detection with An Efficient Framework|

ECCV2022|3D目标检测--点云 |Graph R-CNN: Towards Accurate 3D Object Detection with Semantic-Decorated Local Graph|

ECCV2022|3D目标检测--点云 |CenterFormer: Center-based Transformer for 3D Object Detection|

ECCV2022|点云特征提取--NAS网络 |LidarNAS: Unifying and Searching Neural Architectures for 3D Point Clouds|

ECCV2022|3d目标检测-单目|DEVIANT: Depth EquiVarIAnt NeTwork for Monocular 3D Object Detection|

ECCV2022|3D目标检测-图像|DID-M3D: Decoupling Instance Depth for Monocular 3D Object Detection|

ECCV2022|目标检测(2d)|Should All Proposals be Treated Equally in Object Detection?|

自动驾驶仿真

ECCV2022|仿真环境生成|TRoVE: Transforming Road Scene Datasets into Photorealistic Virtual Environments|

多目标跟踪算法

ECCV2022|目标跟踪--点云 |3D Siamese Transformer Network for Single Object Tracking on Point Clouds|

ECCV2022|3d目标跟踪|PolarMOT: How Far Can Geometric Relations Take Us in 3D Multi-Object Tracking?|

场景分割算法

ECCV2022|点云分割|Online Segmentation of LiDAR Sequences: Dataset and Algorithm|

ECCV2022|全景分割|Panoptic-PartFormer: Learning a Unified model for Panoptic Part Segmentation|

ECCV2022|点云语义分割|Open-world Semantic Segmentation for LIDAR Point Clouds|

ECCV2022|全景分割|Panoptic-PartFormer: Learning a Unified model for Panoptic Part Segmentation|

车道线/可行驶区域

ECCV2022|3d车道线分割--点云|SuperLine3D: Self-supervised Line Segmentation and Description for LiDAR Point Cloud|

ECCV2022|车道线检测|RCLane: Relay Chain Prediction for Lane Detection|

深度估计

ECCV2022|深度估计 |3D-PL: Domain Adaptive Depth Estimation with 3D-aware Pseudo-Labeling|

ECCV2022|深度补全 | GraphCSPN: Geometry-Aware Depth Completion via Dynamic GCNs|

ECCV2022|深度估计(视觉+imu)|Towards Scale-Aware, Robust, and Generalizable Unsupervised Monocular Depth Estimation by Integrating IMU Motion Dynamics|

场景功能

ECCV2022|停车位检测|Towards an Error-free Deep Occupancy Detector for Smart Camera Parking System |

ECCV2022|疲劳检测 |Detecting Driver Drowsiness as an Anomaly Using LSTM Autoencoders

ECCV2022|场景分析--点云|Dynamic 3D Scene Analysis by Point Cloud Accumulation|

ECCV2022|V2X -感知-注意力机制|V2X-ViT: Vehicle-to-Everything Cooperative Perception with Vision Transformer|

相机标定

ECCV2022|相机标定 |Camera Auto-calibration from the Steiner Conic of the Fundamental Matrix|

传感器设计

ECCV2022|环境感知相机 | Overexposure Mask Fusion: Generalizable Reverse ISP Multi-Step Refinement|

【自动驾驶之心】全栈技术交流群

自动驾驶之心是首个自动驾驶开发者社区，聚焦目标检测、语义分割、全景分割、实例分割、关键点检测、车道线、目标跟踪、3D目标检测、BEV感知、多传感器融合、SLAM、光流估计、深度估计、轨迹预测、高精地图、NeRF、规划控制、模型部署落地、自动驾驶仿真测试、硬件配置、AI求职交流等方向；

添加汽车人助理微信邀请入群

备注：学校/公司+方向+昵称