COG: Confidence-aware Optimal Geometric Correspondence for Unsupervised Single-reference Novel Object Pose Estimation

2026-02-28

AI Summary

This work addresses the challenge of establishing robust cross-view correspondences for six-degree-of-freedom pose estimation of novel objects from a single reference image, where occlusion, viewpoint variation, and outliers severely hinder matching. To this end, we propose an unsupervised framework that, for the first time, integrates point-level confidence into the optimal transport formulation to construct differentiable, dense, and robust soft correspondences. By leveraging semantic priors from vision foundation models, our approach enables end-to-end unsupervised pose refinement. Experimental results demonstrate that the unsupervised variant achieves performance on par with current supervised methods, while its supervised counterpart significantly outperforms state-of-the-art techniques.

Technology Category

• Computer Vision: Other Foundations of Computer Vision
• Intelligent Robots: State Estimation
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• 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
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Estimating the 6DoF pose of a novel object with a single reference view is challenging due to occlusions, view-point changes, and outliers. A core difficulty lies in finding robust cross-view correspondences, as existing methods often rely on discrete one-to-one matching that is non-differentiable and tends to collapse onto sparse key-points. We propose Confidence-aware Optimal Geometric Correspondence (COG), an unsupervised framework that formulates correspondence estimation as a confidence-aware optimal transport problem. COG produces balanced soft correspondences by predicting point-wise confidences and injecting them as optimal transport marginals, suppressing non-overlapping regions. Semantic priors from vision foundation models further regularize the correspondences, leading to stable pose estimation. This design integrates confidence into the correspondence finding and pose estimation pipeline, enabling unsupervised learning. Experiments show unsupervised COG achieves comparable performance to supervised methods, and supervised COG outperforms them.

Problem

• 6DoF pose estimation
• novel object
• single-reference
• cross-view correspondence
• unsupervised learning

Innovation

• optimal transport
• confidence-aware correspondence
• unsupervised pose estimation
• soft correspondences
• vision foundation models