Improving Robotic Manipulation with Efficient Geometry-Aware Vision Encoder

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Vuong, An Dinh, Vu, Minh Nhat, Reid, Ian
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909797686706176
author Vuong, An Dinh
Vu, Minh Nhat
Reid, Ian
author_facet Vuong, An Dinh
Vu, Minh Nhat
Reid, Ian
contents Existing RGB-based imitation learning approaches typically employ traditional vision encoders such as ResNet or ViT, which lack explicit 3D reasoning capabilities. Recent geometry-grounded vision models, such as VGGT~\cite{wang2025vggt}, provide robust spatial understanding and are promising candidates to address this limitation. This work investigates the integration of geometry-aware visual representations into robotic manipulation. Our results suggest that incorporating the geometry-aware vision encoder into imitation learning frameworks, including ACT and DP, yields up to 6.5% improvement over standard vision encoders in success rate across single- and bi-manual manipulation tasks in both simulation and real-world settings. Despite these benefits, most geometry-grounded models require high computational cost, limiting their deployment in practical robotic systems. To address this challenge, we propose eVGGT, an efficient geometry-aware encoder distilled from VGGT. eVGGT is nearly 9 times faster and 5 times smaller than VGGT, while preserving strong 3D reasoning capabilities. Code and pretrained models will be released to facilitate further research in geometry-aware robotics.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15880
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Improving Robotic Manipulation with Efficient Geometry-Aware Vision Encoder
Vuong, An Dinh
Vu, Minh Nhat
Reid, Ian
Robotics
Existing RGB-based imitation learning approaches typically employ traditional vision encoders such as ResNet or ViT, which lack explicit 3D reasoning capabilities. Recent geometry-grounded vision models, such as VGGT~\cite{wang2025vggt}, provide robust spatial understanding and are promising candidates to address this limitation. This work investigates the integration of geometry-aware visual representations into robotic manipulation. Our results suggest that incorporating the geometry-aware vision encoder into imitation learning frameworks, including ACT and DP, yields up to 6.5% improvement over standard vision encoders in success rate across single- and bi-manual manipulation tasks in both simulation and real-world settings. Despite these benefits, most geometry-grounded models require high computational cost, limiting their deployment in practical robotic systems. To address this challenge, we propose eVGGT, an efficient geometry-aware encoder distilled from VGGT. eVGGT is nearly 9 times faster and 5 times smaller than VGGT, while preserving strong 3D reasoning capabilities. Code and pretrained models will be released to facilitate further research in geometry-aware robotics.
title Improving Robotic Manipulation with Efficient Geometry-Aware Vision Encoder
topic Robotics
url https://arxiv.org/abs/2509.15880