UniPose9D: Universal Category-Agnostic Object Pose Estimation

2026-07-10

AI Summary

Existing object pose estimation methods exhibit limited generalization to novel categories and unseen scenes. This work proposes a universal, category-agnostic 9D pose estimation approach that jointly predicts object rotation, translation, and metric scale using only instance masks and RGB-D (or RGB with predicted depth) inputs. The method establishes the first unified framework that operates without CAD models, category priors, or reference views. It innovatively integrates point-pair RANSAC with an N-hop Kabsch–Umeyama algorithm and flow matching to resolve symmetry ambiguities, while fusing DINOv2 and PointNet features to regress NOCS coordinates, further refined by adaptive-threshold RANSAC. By aligning multiple datasets to construct a large-scale training set, a single model achieves performance on par with or superior to specialized methods across six benchmarks, demonstrating exceptional cross-category and cross-domain generalization.

Abstract

Object pose estimation is a fundamental problem in 3D vision. Although recent state-of-the-art approaches achieve strong performance, they often overfit to existing benchmarks and exhibit limited generalization to novel categories and unseen scenes. We propose UniPose9D, a category-agnostic foundation model for 9D object pose estimation: given an instance mask/ROI and either an RGB-D observation or an RGB image with predicted depth, the model estimates rotation, translation, and metric size without category labels, CAD models, mean-shape priors, or reference views. Specifically, UniPose9D samples point pairs from the observed object geometry and uses DINOv2 and PointNet features to predict NOCS coordinates for each pair. To improve accuracy, we introduce a point-pair-based RANSAC N-hop Kabsch--Umeyama algorithm with an adaptive threshold. We further employ flow matching to address symmetric ambiguities and construct a large-scale training set by curating and aligning pose annotations from existing public datasets. Experiments across six datasets show that a single unified model can match or surpass specialist methods while generalizing to unseen objects and in-the-wild scenarios. Our code and model are available on https://github.com/qq456cvb/UniPose9D.

Problem

• object pose estimation
• category-agnostic
• generalization
• 3D vision
• novel categories

Innovation

• category-agnostic
• 9D pose estimation
• point-pair sampling
• RANSAC N-hop Kabsch–Umeyama
• flow matching