Cylinder3D: An Effective 3D Framework for Driving-scene LiDAR Semantic Segmentation

by   Hui Zhou, et al.

State-of-the-art methods for large-scale driving-scene LiDAR semantic segmentation often project and process the point clouds in the 2D space. The projection methods includes spherical projection, bird-eye view projection, etc. Although this process makes the point cloud suitable for the 2D CNN-based networks, it inevitably alters and abandons the 3D topology and geometric relations. A straightforward solution to tackle the issue of 3D-to-2D projection is to keep the 3D representation and process the points in the 3D space. In this work, we first perform an in-depth analysis for different representations and backbones in 2D and 3D spaces, and reveal the effectiveness of 3D representations and networks on LiDAR segmentation. Then, we develop a 3D cylinder partition and a 3D cylinder convolution based framework, termed as Cylinder3D, which exploits the 3D topology relations and structures of driving-scene point clouds. Moreover, a dimension-decomposition based context modeling module is introduced to explore the high-rank context information in point clouds in a progressive manner. We evaluate the proposed model on a large-scale driving-scene dataset, i.e. SematicKITTI. Our method achieves state-of-the-art performance and outperforms existing methods by 6 mIoU.


page 1

page 3

page 5

page 6

page 7

page 8

page 9

page 10


Cylindrical and Asymmetrical 3D Convolution Networks for LiDAR Segmentation

State-of-the-art methods for large-scale driving-scene LiDAR segmentatio...

PCB-RandNet: Rethinking Random Sampling for LIDAR Semantic Segmentation in Autonomous Driving Scene

Fast and efficient semantic segmentation of large-scale LiDAR point clou...

3DContextNet: K-d Tree Guided Hierarchical Learning of Point Clouds Using Local Contextual Cues

3D data such as point clouds and meshes are becoming more and more avail...

Learning Inner-Group Relations on Point Clouds

The prevalence of relation networks in computer vision is in stark contr...

Weakly Supervised Semantic Segmentation in 3D Graph-Structured Point Clouds of Wild Scenes

The deficiency of 3D segmentation labels is one of the main obstacles to...

Bidirectional Projection Network for Cross Dimension Scene Understanding

2D image representations are in regular grids and can be processed effic...

LAPTNet-FPN: Multi-scale LiDAR-aided Projective Transform Network for Real Time Semantic Grid Prediction

Semantic grids can be useful representations of the scene around an auto...

Please sign up or login with your details

Forgot password? Click here to reset