GBlobs: Local LiDAR Geometry for Improved Sensor Placement Generalization

TL;DR

This paper introduces GBlobs, a local point cloud feature descriptor that improves 3D object detection models' ability to generalize across different LiDAR sensor placements by avoiding reliance on absolute position data.

Contribution

The paper presents GBlobs, a novel local feature descriptor that enhances model generalization across diverse LiDAR configurations by mitigating geometric shortcuts.

Findings

Achieved state-of-the-art performance on RoboSense 2025 challenge

Enhanced model robustness to sensor placement variations

Circumvented geometric shortcuts in 3D detection models

Abstract

This technical report outlines the top-ranking solution for RoboSense 2025: Track 3, achieving state-of-the-art performance on 3D object detection under various sensor placements. Our submission utilizes GBlobs, a local point cloud feature descriptor specifically designed to enhance model generalization across diverse LiDAR configurations. Current LiDAR-based 3D detectors often suffer from a \enquote{geometric shortcut} when trained on conventional global features (\ie, absolute Cartesian coordinates). This introduces a position bias that causes models to primarily rely on absolute object position rather than distinguishing shape and appearance characteristics. Although effective for in-domain data, this shortcut severely limits generalization when encountering different point distributions, such as those resulting from varying sensor placements. By using GBlobs as network input features, we effectively circumvent this geometric shortcut, compelling the network to learn robust, object-centric representations. This approach significantly enhances the model's ability to generalize, resulting in the exceptional performance demonstrated in this challenge.