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

2025-07-24

AI Summary

Existing 3D point cloud test-time adaptation (TTA) methods struggle to dynamically adapt under domain shifts due to the inherent disorder and unstructured nature of point clouds. Method: This paper pioneers the migration of TTA to the spectral graph domain: it constructs an outlier-aware adjacency graph over point clouds and applies Graph Fourier Transform (GFT) to extract global structural representations; only the lowest 10% frequency components are optimized for parameter-efficient adaptation; and a feature-map-guided spectral self-training mechanism jointly refines both the model and its spectral representation. Contribution/Results: The proposed method significantly outperforms state-of-the-art 3D TTA approaches across multiple benchmarks, achieving substantial gains in classification accuracy while reducing computational overhead. These results validate the effectiveness and efficiency of sparse spectral-domain optimization for point cloud TTA.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: 3D Computer Vision
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

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.

Problem

• Adapting pre-trained models to 3D point cloud domain shifts efficiently
• Overcoming irregular point cloud structure with graph spectral methods
• Reducing computational cost by optimizing key frequency components

Innovation

• Shifts adaptation to graph spectral domain
• Optimizes lowest 10% frequency components
• Uses eigenmap-guided self-training strategy