Reconfigurable Voxels: A New Representation for LiDAR-Based Point Clouds
arXiv:2004.02724
Abstract
LiDAR is an important method for autonomous driving systems to sense the environment. The point clouds obtained by LiDAR typically exhibit sparse and irregular distribution, thus posing great challenges to the detection of 3D objects, especially those that are small and distant. To tackle this difficulty, we propose Reconfigurable Voxels, a new approach to constructing representations from 3D point clouds. Specifically, we devise a biased random walk scheme, which adaptively covers each neighborhood with a fixed number of voxels based on the local spatial distribution and produces a representation by integrating the points in the chosen neighbors. We found empirically that this approach effectively improves the stability of voxel features, especially for sparse regions. Experimental results on multiple benchmarks, including nuScenes, Lyft, and KITTI, show that this new representation can remarkably improve the detection performance for small and distant objects, without incurring noticeable overhead costs.
Conference on Robot Learning (CoRL) 2020
References in corpus (5)
- Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection
- Frustum ConvNet: Sliding Frustums to Aggregate Local Point-Wise Features for Amodal 3D Object Detection
- From Points to Parts: 3D Object Detection from Point Cloud with Part-aware and Part-aggregation Network
- Deformable Kernels: Adapting Effective Receptive Fields for Object Deformation
- Deformable Filter Convolution for Point Cloud Reasoning
Cited by in corpus (7)
- Cylinder3D: An Effective 3D Framework for Driving-scene LiDAR Semantic Segmentation
- Probabilistic and Geometric Depth: Detecting Objects in Perspective
- Graph Neural Network and Spatiotemporal Transformer Attention for 3D Video Object Detection from Point Clouds
- Density-aware Chamfer Distance as a Comprehensive Metric for Point Cloud Completion
- FCOS3D: Fully Convolutional One-Stage Monocular 3D Object Detection
- Channel-wise Alignment for Adaptive Object Detection
- Input-Output Balanced Framework for Long-tailed LiDAR Semantic Segmentation