Deformable Filter Convolution for Point Cloud Reasoning
arXiv:1907.13079
Abstract
Point clouds are the native output of many real-world 3D sensors. To borrow the success of 2D convolutional network architectures, a majority of popular 3D perception models voxelize the points, which can result in a loss of local geometric details that cannot be recovered. In this paper, we propose a novel learnable convolution layer for processing 3D point cloud data directly. Instead of discretizing points into fixed voxels, we deform our learnable 3D filters to match with the point cloud shape. We propose to combine voxelized backbone networks with our deformable filter layer at 1) the network input stream and 2) the output prediction layers to enhance point level reasoning. We obtain state-of-the-art results on LiDAR semantic segmentation and producing a significant gain in performance on LiDAR object detection.
References in corpus (6)
- Semi-Supervised Classification with Graph Convolutional Networks
- Deep Continuous Fusion for Multi-Sensor 3D Object Detection
- PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation
- HDNET: Exploiting HD Maps for 3D Object Detection
- SO-Net: Self-Organizing Network for Point Cloud Analysis
- SPLATNet: Sparse Lattice Networks for Point Cloud Processing
Cited by in corpus (7)
- Learning Semantic Segmentation of Large-Scale Point Clouds with Random Sampling
- Rotation-Invariant Local-to-Global Representation Learning for 3D Point Cloud
- Deformable Kernels: Adapting Effective Receptive Fields for Object Deformation
- Reconfigurable Voxels: A New Representation for LiDAR-Based Point Clouds
- Intrinsic-Extrinsic Convolution and Pooling for Learning on 3D Protein Structures
- StrObe: Streaming Object Detection from LiDAR Packets
- Pointwise Attention-Based Atrous Convolutional Neural Networks