SemanticVoxels: Sequential Fusion for 3D Pedestrian Detection using LiDAR Point Cloud and Semantic Segmentation

TL;DR

SemanticVoxels introduces a multi-level fusion method combining LiDAR and semantic segmentation to improve 3D pedestrian detection, outperforming existing approaches on the KITTI benchmark.

Contribution

It generalizes PointPainting for multi-level fusion of semantic and geometric data, enhancing detection accuracy in challenging scenarios.

Findings

Achieves state-of-the-art results on KITTI for 3D and BEV pedestrian detection.

Outperforms current methods in detecting challenging pedestrian cases.

Effective fusion at different levels improves detection robustness.

Abstract

3D pedestrian detection is a challenging task in automated driving because pedestrians are relatively small, frequently occluded and easily confused with narrow vertical objects. LiDAR and camera are two commonly used sensor modalities for this task, which should provide complementary information. Unexpectedly, LiDAR-only detection methods tend to outperform multisensor fusion methods in public benchmarks. Recently, PointPainting has been presented to eliminate this performance drop by effectively fusing the output of a semantic segmentation network instead of the raw image information. In this paper, we propose a generalization of PointPainting to be able to apply fusion at different levels. After the semantic augmentation of the point cloud, we encode raw point data in pillars to get geometric features and semantic point data in voxels to get semantic features and fuse them in an effective way. Experimental results on the KITTI test set show that SemanticVoxels achieves state-of-the-art performance in both 3D and bird's eye view pedestrian detection benchmarks. In particular, our approach demonstrates its strength in detecting challenging pedestrian cases and outperforms current state-of-the-art approaches.