PSA-Det3D: Pillar Set Abstraction for 3D object Detection

TL;DR

This paper introduces PSA-Det3D, a novel 3D object detection method that enhances small object detection in point clouds through a pillar set abstraction and foreground point compensation, achieving superior accuracy on KITTI.

Contribution

The paper proposes a new pillar set abstraction and foreground point compensation technique to improve small object detection in 3D point clouds.

Findings

Outperforms existing methods on KITTI benchmark

Significantly improves detection of small objects

Effective aggregation of point-wise features

Abstract

Small object detection for 3D point cloud is a challenging problem because of two limitations: (1) Perceiving small objects is much more diffcult than normal objects due to the lack of valid points. (2) Small objects are easily blocked which breaks the shape of their meshes in 3D point cloud. In this paper, we propose a pillar set abstraction (PSA) and foreground point compensation (FPC) and design a point-based detection network, PSA-Det3D, to improve the detection performance for small object. The PSA embeds a pillar query operation on the basis of set abstraction (SA) to expand its receptive field of the network, which can aggregate point-wise features effectively. To locate more occluded objects, we persent a proposal generation layer consisting of a foreground point segmentation and a FPC module. Both the foreground points and the estimated centers are finally fused together to generate the detection result. The experiments on the KITTI 3D detection benchmark show that our proposed PSA-Det3D outperforms other algorithms with high accuracy for small object detection.