New PANY framework achieves state-of-the-art object pose estimation

Researchers have developed PANY, a novel model-free framework for estimating the 6D pose of unseen objects, designed for open-world robotics and embodied perception. Unlike previous methods limited to pairwise matching, PANY utilizes a multi-view transformer geometry backbone to learn view-consistent geometry and cross-view alignment cues, enabling robust performance even with limited query-reference overlap and occlusion. The framework supports both RGB and RGB-D inputs and can leverage sparse reference views or additional unposed assist views for improved geometric coverage and pose accuracy. Experiments demonstrate PANY achieves state-of-the-art results, outperforming existing model-free approaches by significant margins on benchmarks like YCB-V and LM-O. AI

IMPACT This new framework could significantly advance robotics and embodied perception by enabling more robust object recognition and manipulation in complex environments.

RANK_REASON The item is an academic paper detailing a new method for object pose estimation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• Litmaps
• ScienceCast
• scite Smart Citations
• YCB-V

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →