PQDT: Pseudo-Query Dual Transformer for Robust Point Cloud Restoration

2026-05-24

AI Summary

Real-world point clouds often suffer from various degradations—such as missing regions, noise, outliers, and non-uniform density—due to sensor limitations or occlusions, severely compromising 3D perception performance. To address this challenge, this work proposes a unified point cloud restoration network that takes only the raw point cloud as input and leverages a novel pseudo-query dual-Transformer architecture to decompose geometric reconstruction into two synergistic stages, simultaneously preserving fine local details and enhancing structural clarity and robustness. Notably, the method operates without any auxiliary information and achieves state-of-the-art performance across multiple benchmarks. It effectively handles complex combinations of degradation types—including completion, deformation correction, and denoising—enabling high-quality, adaptive point cloud restoration.

Abstract

Point clouds are a fundamental 3D representation in computer vision, enabling a wide range of perception tasks. However, real-world point clouds often suffer from degradations such as incompleteness, noise, outliers, and irregular density, caused by sensor limitations or occlusions. Recovering clean and detailed shapes from such degraded data is crucial for downstream applications. While existing learning-based methods achieve progress on individual tasks like completion or denoising, they typically rely on global bottleneck features, which lose fine-grained geometry and remain sensitive to varying input quality. We propose a unified 3D restoration network that directly takes point clouds as input and adaptively reconstructs high-quality geometry under diverse degradation scenarios. At the core of our approach is a Pseudo-Query module, implemented within a Transformer backbone, which reformulates geometric translation into two cooperative stages to enhance structural clarity, robustness, and local detail preservation. Extensive experiments on curated benchmarks demonstrate that our approach surpasses state-of-the-art performance in general 3D restoration. It effectively handles complex combinations of completion, deformation, and denoising degradations. With this work, we provide a novel unified, point-only backbone for robust 3D restoration, enabling more versatile 3D perception.

Problem

• point cloud restoration
• 3D degradation
• noise and outliers
• incomplete data
• irregular density

Innovation

• Pseudo-Query
• Dual Transformer
• Point Cloud Restoration
• Robust 3D Reconstruction
• Unified Architecture