StarNet: Targeted Computation for Object Detection in Point Clouds

Abstract

Detecting objects from LiDAR point clouds is an important component of self-driving car technology as LiDAR provides high resolution spatial information. Previous work on point-cloud 3D object detection has re-purposed convolutional approaches from traditional camera imagery. In this work, we present an object detection system called StarNet designed specifically to take advantage of the sparse and 3D nature of point cloud data. StarNet is entirely point-based, uses no global information, has data dependent anchors, and uses sampling instead of learned region proposals. We demonstrate how this design leads to competitive or superior performance on the large Waymo Open Dataset and the KITTI detection dataset, as compared to convolutional baselines. In particular, we show how our detector can outperform a competitive baseline on Pedestrian detection on the Waymo Open Dataset by more than 7 absolute mAP while being more computationally efficient. We show how our redesign---namely using only local information and using sampling instead of learned proposals---leads to a significantly more flexible and adaptable system: we demonstrate how we can vary the computational cost of a single trained StarNet without retraining, and how we can target proposals towards areas of interest with priors and heuristics. Finally, we show how our design allows for incorporating temporal context by using detections from previous frames to target computation of the detector, which leads to further improvements in performance without additional computational cost.

References in corpus (8)

• Adam: A Method for Stochastic Optimization
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• End to End Learning for Self-Driving Cars
• How Powerful are Graph Neural Networks?
• Fast and Furious: Real Time End-to-End 3D Detection, Tracking and Motion Forecasting with a Single Convolutional Net
• HDNET: Exploiting HD Maps for 3D Object Detection
• IPOD: Intensive Point-based Object Detector for Point Cloud
• PointConv: Deep Convolutional Networks on 3D Point Clouds

Cited by in corpus (27)

• Deep Multi-modal Object Detection and Semantic Segmentation for Autonomous Driving: Datasets, Methods, and Challenges
• PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection
• AFDet: Anchor Free One Stage 3D Object Detection
• Center-based 3D Object Detection and Tracking
• Affinity and Diversity: Quantifying Mechanisms of Data Augmentation
• Autonomous Driving with Deep Learning: A Survey of State-of-Art Technologies
• Pillar-based Object Detection for Autonomous Driving
• PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection
• PseudoAugment: Learning to Use Unlabeled Data for Data Augmentation in Point Clouds
• Improving 3D Object Detection through Progressive Population Based Augmentation
• Range Conditioned Dilated Convolutions for Scale Invariant 3D Object Detection
• Improving 3D Object Detection with Channel-wise Transformer
• An LSTM Approach to Temporal 3D Object Detection in LiDAR Point Clouds
• Pyramid R-CNN: Towards Better Performance and Adaptability for 3D Object Detection
• STINet: Spatio-Temporal-Interactive Network for Pedestrian Detection and Trajectory Prediction
• Multi-Frame to Single-Frame: Knowledge Distillation for 3D Object Detection
• RSN: Range Sparse Net for Efficient, Accurate LiDAR 3D Object Detection
• Temporal-Channel Transformer for 3D Lidar-Based Video Object Detection in Autonomous Driving
• Streaming Object Detection for 3-D Point Clouds
• Voxel Transformer for 3D Object Detection
• Pseudo-labeling for Scalable 3D Object Detection
• 3D-MAN: 3D Multi-frame Attention Network for Object Detection
• LiDAR R-CNN: An Efficient and Universal 3D Object Detector
• 3D-FCT: Simultaneous 3D Object Detection and Tracking Using Feature Correlation
• Improved Pillar with Fine-grained Feature for 3D Object Detection
• Structure Information is the Key: Self-Attention RoI Feature Extractor in 3D Object Detection
• To the Point: Efficient 3D Object Detection in the Range Image with Graph Convolution Kernels