CL3D: Unsupervised Domain Adaptation for Cross-LiDAR 3D Detection

TL;DR

CL3D introduces an unsupervised domain adaptation approach for cross-LiDAR 3D detection, leveraging geometric and motion invariance to improve detection across different LiDAR sensors and environments.

Contribution

The paper proposes two novel modules for geometric and motion alignment, enhancing self-training for unsupervised domain adaptation in 3D detection.

Findings

Achieves state-of-the-art results on cross-device datasets.

Effective in scenarios with large gaps between LiDAR types.

Utilizes geometric and motion features for domain invariance.

Abstract

Domain adaptation for Cross-LiDAR 3D detection is challenging due to the large gap on the raw data representation with disparate point densities and point arrangements. By exploring domain-invariant 3D geometric characteristics and motion patterns, we present an unsupervised domain adaptation method that overcomes above difficulties. First, we propose the Spatial Geometry Alignment module to extract similar 3D shape geometric features of the same object class to align two domains, while eliminating the effect of distinct point distributions. Second, we present Temporal Motion Alignment module to utilize motion features in sequential frames of data to match two domains. Prototypes generated from two modules are incorporated into the pseudo-label reweighting procedure and contribute to our effective self-training framework for the target domain. Extensive experiments show that our method achieves state-of-the-art performance on cross-device datasets, especially for the datasets with large gaps captured by mechanical scanning LiDARs and solid-state LiDARs in various scenes. Project homepage is at https://github.com/4DVLab/CL3D.git