Any6D: Model-free 6D Pose Estimation of Novel Objects

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lee, Taeyeop, Wen, Bowen, Kang, Minjun, Kang, Gyuree, Kweon, In So, Yoon, Kuk-Jin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913756440690688
author Lee, Taeyeop
Wen, Bowen
Kang, Minjun
Kang, Gyuree
Kweon, In So
Yoon, Kuk-Jin
author_facet Lee, Taeyeop
Wen, Bowen
Kang, Minjun
Kang, Gyuree
Kweon, In So
Yoon, Kuk-Jin
contents We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
format Preprint
id arxiv_https___arxiv_org_abs_2503_18673
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Any6D: Model-free 6D Pose Estimation of Novel Objects
Lee, Taeyeop
Wen, Bowen
Kang, Minjun
Kang, Gyuree
Kweon, In So
Yoon, Kuk-Jin
Computer Vision and Pattern Recognition
Artificial Intelligence
Robotics
We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
title Any6D: Model-free 6D Pose Estimation of Novel Objects
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Robotics
url https://arxiv.org/abs/2503.18673