Searching Efficient 3D Architectures with Sparse Point-Voxel Convolution

TL;DR

This paper introduces Sparse Point-Voxel Convolution (SPVConv), a lightweight 3D module that enhances 3D perception models for autonomous driving by preserving fine details and enabling efficient neural architecture search, leading to state-of-the-art results.

Contribution

The paper proposes SPVConv, a novel 3D convolution module combining point-based and voxel-based methods, and introduces 3D-NAS for optimizing 3D network architectures.

Findings

SPVNAS outperforms MinkowskiNet by 3.3% on SemanticKITTI.

Achieves 8x less computation and 3x faster inference with higher accuracy.

Improves 3D object detection performance on KITTI dataset.

Abstract

Self-driving cars need to understand 3D scenes efficiently and accurately in order to drive safely. Given the limited hardware resources, existing 3D perception models are not able to recognize small instances (e.g., pedestrians, cyclists) very well due to the low-resolution voxelization and aggressive downsampling. To this end, we propose Sparse Point-Voxel Convolution (SPVConv), a lightweight 3D module that equips the vanilla Sparse Convolution with the high-resolution point-based branch. With negligible overhead, this point-based branch is able to preserve the fine details even from large outdoor scenes. To explore the spectrum of efficient 3D models, we first define a flexible architecture design space based on SPVConv, and we then present 3D Neural Architecture Search (3D-NAS) to search the optimal network architecture over this diverse design space efficiently and effectively. Experimental results validate that the resulting SPVNAS model is fast and accurate: it outperforms the state-of-the-art MinkowskiNet by 3.3%, ranking 1st on the competitive SemanticKITTI leaderboard. It also achieves 8x computation reduction and 3x measured speedup over MinkowskiNet with higher accuracy. Finally, we transfer our method to 3D object detection, and it achieves consistent improvements over the one-stage detection baseline on KITTI.