THE-Pose: Topological Prior with Hybrid Graph Fusion for Estimating Category-Level 6D Object Pose

2025-12-10

AI Summary

To address the insufficient robustness in category-level 6D object pose estimation—caused by intra-class shape deformation, occlusion, and visual ambiguity—this paper introduces, for the first time, surface embeddings to explicitly encode topological priors. We propose an Adaptive Hybrid Graph Fusion (HGF) mechanism that jointly models 2D semantic consistency and 3D geometric structure. Our method integrates surface topology embeddings, image features, point cloud geometric encodings, and a learnable graph fusion module to enable cross-modal structured reasoning. Evaluated on the REAL275 benchmark, our approach achieves a 35.8% improvement over the 3D-GC baseline HS-Pose and surpasses prior state-of-the-art methods by 7.2% in average precision. It demonstrates significantly enhanced generalization capability and pose estimation accuracy under challenging real-world conditions, including severe occlusion and large intra-class variations.

Technology Category

• Computer Vision: 3D Computer Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Graph-based Machine Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for 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: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Category-level object pose estimation requires both global context and local structure to ensure robustness against intra-class variations. However, 3D graph convolution (3D-GC) methods only focus on local geometry and depth information, making them vulnerable to complex objects and visual ambiguities. To address this, we present THE-Pose, a novel category-level 6D pose estimation framework that leverages a topological prior via surface embedding and hybrid graph fusion. Specifically, we extract consistent and invariant topological features from the image domain, effectively overcoming the limitations inherent in existing 3D-GC based methods. Our Hybrid Graph Fusion (HGF) module adaptively integrates the topological features with point-cloud features, seamlessly bridging 2D image context and 3D geometric structure. These fused features ensure stability for unseen or complicated objects, even under significant occlusions. Extensive experiments on the REAL275 dataset show that THE-Pose achieves a 35.8% improvement over the 3D-GC baseline (HS-Pose) and surpasses the previous state-of-the-art by 7.2% across all key metrics. The code is avaialbe on https://github.com/EHxxx/THE-Pose

Problem

• Estimates 6D object poses for categories with intra-class variations
• Overcomes limitations of 3D graph convolution methods using topological priors
• Integrates 2D image context with 3D geometry via hybrid graph fusion

Innovation

• Leverages topological prior via surface embedding and hybrid graph fusion
• Integrates topological features with point-cloud features adaptively
• Bridges 2D image context and 3D geometric structure seamlessly