NUC-Net: Non-uniform Cylindrical Partition Network for Efficient LiDAR Semantic Segmentation

TL;DR

NUC-Net introduces a non-uniform cylindrical partitioning approach for LiDAR semantic segmentation that enhances accuracy and efficiency, reducing computational costs and better handling unbalanced point distributions.

Contribution

The paper proposes the API method for non-uniform partitioning and a multi-scale aggregation technique, significantly improving LiDAR segmentation performance and efficiency over uniform methods.

Findings

Achieves state-of-the-art results on SemanticKITTI and nuScenes datasets.

Fours times faster training and two times less GPU memory usage.

Three times faster inference speed.

Abstract

LiDAR semantic segmentation plays a vital role in autonomous driving. Existing voxel-based methods for LiDAR semantic segmentation apply uniform partition to the 3D LiDAR point cloud to form a structured representation based on cartesian/cylindrical coordinates. Although these methods show impressive performance, the drawback of existing voxel-based methods remains in two aspects: (1) it requires a large enough input voxel resolution, which brings a large amount of computation cost and memory consumption. (2) it does not well handle the unbalanced point distribution of LiDAR point cloud. In this paper, we propose a non-uniform cylindrical partition network named NUC-Net to tackle the above challenges. Specifically, we propose the Arithmetic Progression of Interval (API) method to non-uniformly partition the radial axis and generate the voxel representation which is representative and efficient. Moreover, we propose a non-uniform multi-scale aggregation method to improve contextual information. Our method achieves state-of-the-art performance on SemanticKITTI and nuScenes datasets with much faster speed and much less training time. And our method can be a general component for LiDAR semantic segmentation, which significantly improves both the accuracy and efficiency of the uniform counterpart by $4 \times$ training faster and $2 \times$ GPU memory reduction and $3 \times$ inference speedup. We further provide theoretical analysis towards understanding why NUC is effective and how point distribution affects performance. Code is available at \href{https://github.com/alanWXZ/NUC-Net}{https://github.com/alanWXZ/NUC-Net}.