Enhancing Rotation-Invariant 3D Learning with Global Pose Awareness and Attention Mechanisms

2025-11-11

AI Summary

Existing rotation-invariant (RI) 3D point cloud methods rely on handcrafted RI features, leading to loss of global pose information and poor discrimination of symmetric structures (e.g., aircraft wings) or spatially similar parts. To address this, we propose Shadow-informed Pose Features and RIAttnConv—a novel mechanism that dynamically learns the optimal global rotation via the Bingham distribution parameterized by unit quaternions, integrated with attention-driven feature aggregation and a learnable rotation-alignment module. This is the first approach unifying RI representation with explicit global pose awareness. It effectively mitigates local feature collapse and significantly enhances fine-grained discrimination of symmetric geometries. Extensive experiments demonstrate state-of-the-art performance on 3D classification and part segmentation under arbitrary rotations, with particularly pronounced gains in complex symmetric scenarios.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Feature Construction/Reformulation
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Recent advances in rotation-invariant (RI) learning for 3D point clouds typically replace raw coordinates with handcrafted RI features to ensure robustness under arbitrary rotations. However, these approaches often suffer from the loss of global pose information, making them incapable of distinguishing geometrically similar but spatially distinct structures. We identify that this limitation stems from the restricted receptive field in existing RI methods, leading to Wing-tip feature collapse, a failure to differentiate symmetric components (e.g., left and right airplane wings) due to indistinguishable local geometries. To overcome this challenge, we introduce the Shadow-informed Pose Feature (SiPF), which augments local RI descriptors with a globally consistent reference point (referred to as the'shadow') derived from a learned shared rotation. This mechanism enables the model to preserve global pose awareness while maintaining rotation invariance. We further propose Rotation-invariant Attention Convolution (RIAttnConv), an attention-based operator that integrates SiPFs into the feature aggregation process, thereby enhancing the model's capacity to distinguish structurally similar components. Additionally, we design a task-adaptive shadow locating module based on the Bingham distribution over unit quaternions, which dynamically learns the optimal global rotation for constructing consistent shadows. Extensive experiments on 3D classification and part segmentation benchmarks demonstrate that our approach substantially outperforms existing RI methods, particularly in tasks requiring fine-grained spatial discrimination under arbitrary rotations.

Problem

• Distinguishing geometrically similar but spatially distinct 3D structures under rotation
• Overcoming wing-tip feature collapse in symmetric components with indistinguishable local geometries
• Preserving global pose awareness while maintaining rotation invariance in 3D learning

Innovation

• Shadow-informed Pose Feature enhances global pose awareness
• Rotation-invariant Attention Convolution integrates attention mechanisms
• Task-adaptive shadow locating module learns optimal global rotation