3D Test-time Adaptation via Graph Spectral Driven Point Shift

TL;DR

This paper introduces GSDTTA, a novel 3D point cloud test-time adaptation method that leverages graph spectral domain transformations to efficiently adapt models to new domains, outperforming existing methods.

Contribution

The paper proposes a spectral domain adaptation approach for 3D point clouds, reducing computational costs and requiring fewer parameters compared to spatial-domain methods.

Findings

Outperforms existing 3D TTA methods on benchmark datasets.

Efficient adaptation by optimizing only the lowest 10% of spectral components.

Effective spectral-driven point shift improves classification accuracy.

Abstract

While test-time adaptation (TTA) methods effectively address domain shifts by dynamically adapting pre-trained models to target domain data during online inference, their application to 3D point clouds is hindered by their irregular and unordered structure. Current 3D TTA methods often rely on computationally expensive spatial-domain optimizations and may require additional training data. In contrast, we propose Graph Spectral Domain Test-Time Adaptation (GSDTTA), a novel approach for 3D point cloud classification that shifts adaptation to the graph spectral domain, enabling more efficient adaptation by capturing global structural properties with fewer parameters. Point clouds in target domain are represented as outlier-aware graphs and transformed into graph spectral domain by Graph Fourier Transform (GFT). For efficiency, adaptation is performed by optimizing only the lowest 10% of frequency components, which capture the majority of the point cloud's energy. An inverse GFT (IGFT) is then applied to reconstruct the adapted point cloud with the graph spectral-driven point shift. This process is enhanced by an eigenmap-guided self-training strategy that iteratively refines both the spectral adjustments and the model parameters. Experimental results and ablation studies on benchmark datasets demonstrate the effectiveness of GSDTTA, outperforming existing TTA methods for 3D point cloud classification.