Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection
arXiv:1908.09492
Abstract
This report presents our method which wins the nuScenes3D Detection Challenge [17] held in Workshop on Autonomous Driving(WAD, CVPR 2019). Generally, we utilize sparse 3D convolution to extract rich semantic features, which are then fed into a class-balanced multi-head network to perform 3D object detection. To handle the severe class imbalance problem inherent in the autonomous driving scenarios, we design a class-balanced sampling and augmentation strategy to generate a more balanced data distribution. Furthermore, we propose a balanced group-ing head to boost the performance for the categories withsimilar shapes. Based on the Challenge results, our methodoutperforms the PointPillars [14] baseline by a large mar-gin across all metrics, achieving state-of-the-art detection performance on the nuScenes dataset. Code will be released at CBGS.
technical report
References in corpus (8)
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
- Open3D: A Modern Library for 3D Data Processing
- Deep Continuous Fusion for Multi-Sensor 3D Object Detection
- Submanifold Sparse Convolutional Networks
- Multi-Task Learning as Multi-Objective Optimization
- Multi-View 3D Object Detection Network for Autonomous Driving
- STD: Sparse-to-Dense 3D Object Detector for Point Cloud
- PointFusion: Deep Sensor Fusion for 3D Bounding Box Estimation
Cited by in corpus (42)
- PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection
- A Survey on Approximate Edge AI for Energy Efficient Autonomous Driving Services
- AFDet: Anchor Free One Stage 3D Object Detection
- Graph Neural Network and Spatiotemporal Transformer Attention for 3D Video Object Detection from Point Clouds
- Center-based 3D Object Detection and Tracking
- DEFT: Detection Embeddings for Tracking
- Fully Sparse Fusion for 3D Object Detection
- Probabilistic 3D Multi-Object Tracking for Autonomous Driving
- 3D Multi-Object Tracking: A Baseline and New Evaluation Metrics
- Autonomous Driving with Deep Learning: A Survey of State-of-Art Technologies
- FCOS3D: Fully Convolutional One-Stage Monocular 3D Object Detection
- Exploring Data Augmentation for Multi-Modality 3D Object Detection
- Reconfigurable Voxels: A New Representation for LiDAR-Based Point Clouds
- Object as Hotspots: An Anchor-Free 3D Object Detection Approach via Firing of Hotspots
- Tracking Objects as Points
- Anytime-Lidar: Deadline-aware 3D Object Detection
- PTP: Parallelized Tracking and Prediction with Graph Neural Networks and Diversity Sampling
- LiDAR-based Online 3D Video Object Detection with Graph-based Message Passing and Spatiotemporal Transformer Attention
- DSDNet: Deep Structured self-Driving Network
- RadarNet: Exploiting Radar for Robust Perception of Dynamic Objects
- Active Perception using Light Curtains for Autonomous Driving
- SSN: Shape Signature Networks for Multi-class Object Detection from Point Clouds
- FusionPainting: Multimodal Fusion with Adaptive Attention for 3D Object Detection
- 3D Object Detection with Pointformer
- PnPNet: End-to-End Perception and Prediction with Tracking in the Loop
- 3D for Free: Crossmodal Transfer Learning using HD Maps
- It's All Around You: Range-Guided Cylindrical Network for 3D Object Detection
- An Overview Of 3D Object Detection
- The efficacy of Neural Planning Metrics: A meta-analysis of PKL on nuScenes
- Temporal-Channel Transformer for 3D Lidar-Based Video Object Detection in Autonomous Driving
- Real-Time Anchor-Free Single-Stage 3D Detection with IoU-Awareness
- VIN: Voxel-based Implicit Network for Joint 3D Object Detection and Segmentation for Lidars
- Safety-Oriented Pedestrian Motion and Scene Occupancy Forecasting
- MBDF-Net: Multi-Branch Deep Fusion Network for 3D Object Detection
- PanoNet3D: Combining Semantic and Geometric Understanding for LiDARPoint Cloud Detection
- AutoSelect: Automatic and Dynamic Detection Selection for 3D Multi-Object Tracking
- DR-SPAAM: A Spatial-Attention and Auto-regressive Model for Person Detection in 2D Range Data
- A two-stage data association approach for 3D Multi-object Tracking
- Uncertainty-aware Self-supervised 3D Data Association
- Torch-Points3D: A Modular Multi-Task Frameworkfor Reproducible Deep Learning on 3D Point Clouds
- Real-time 3D Object Detection using Feature Map Flow
- Comparative study of subset selection methods for rapid prototyping of 3D object detection algorithms