Learning Shape-Independent Transformation via Spherical Representations for Category-Level Object Pose Estimation

2025-03-18

AI Summary

Category-level object pose estimation suffers from semantic inconsistency across objects due to reliance on canonical shape-specific coordinate systems. To address this, we propose SpherePose—a shape-agnostic pose regression framework that employs a unit sphere as a universal proxy geometry. Our method introduces SO(3)-invariant point feature encoding, spherical anchor-based attention, and a hyperbolic-space correspondence loss to enable robust cross-shape point matching. The framework is fully end-to-end trainable and effectively mitigates performance degradation under irregular or sparse point cloud inputs. Evaluated on CAMERA25, REAL275, and HouseCat6D benchmarks, SpherePose achieves state-of-the-art accuracy in 6D pose estimation. Moreover, it demonstrates significantly improved generalization across unseen object categories and enhanced robustness to geometric variations and partial observations—setting new standards for category-level pose estimation.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Intelligent Robots: State Estimation
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Category-level object pose estimation aims to determine the pose and size of novel objects in specific categories. Existing correspondence-based approaches typically adopt point-based representations to establish the correspondences between primitive observed points and normalized object coordinates. However, due to the inherent shape-dependence of canonical coordinates, these methods suffer from semantic incoherence across diverse object shapes. To resolve this issue, we innovatively leverage the sphere as a shared proxy shape of objects to learn shape-independent transformation via spherical representations. Based on this insight, we introduce a novel architecture called SpherePose, which yields precise correspondence prediction through three core designs. Firstly, We endow the point-wise feature extraction with SO(3)-invariance, which facilitates robust mapping between camera coordinate space and object coordinate space regardless of rotation transformation. Secondly, the spherical attention mechanism is designed to propagate and integrate features among spherical anchors from a comprehensive perspective, thus mitigating the interference of noise and incomplete point cloud. Lastly, a hyperbolic correspondence loss function is designed to distinguish subtle distinctions, which can promote the precision of correspondence prediction. Experimental results on CAMERA25, REAL275 and HouseCat6D benchmarks demonstrate the superior performance of our method, verifying the effectiveness of spherical representations and architectural innovations.

Problem

• Address shape-dependence in object pose estimation
• Improve semantic coherence across diverse object shapes
• Enhance precision in correspondence prediction using spherical representations

Innovation

• Uses spherical representations for shape-independent transformation.
• Introduces SO(3)-invariant point-wise feature extraction.
• Implements spherical attention mechanism for feature integration.