VoxAfford: Multi-Scale Voxel-Token Fusion for Open-Vocabulary 3D Affordance Detection

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Sun, Haowen, Zhang, Shaolong, Li, Mingyang, Ma, Chengzhong, Chen, Xinzhe, Cui, Qiongjie, Chen, Xingyu, Liu, Zeyang, Lan, Xuguang
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915974179979264
author Sun, Haowen
Zhang, Shaolong
Li, Mingyang
Ma, Chengzhong
Chen, Xinzhe
Cui, Qiongjie
Chen, Xingyu
Liu, Zeyang
Lan, Xuguang
author_facet Sun, Haowen
Zhang, Shaolong
Li, Mingyang
Ma, Chengzhong
Chen, Xinzhe
Cui, Qiongjie
Chen, Xingyu
Liu, Zeyang
Lan, Xuguang
contents Open-vocabulary 3D affordance detection requires localizing interaction regions on point clouds given novel affordance descriptions. Recent methods extend multimodal large language models (MLLMs) with special output tokens that are decoded into segmentation masks. However, these tokens are produced through autoregressive generation, which models sequential dependencies rather than spatial neighborhood relations, leaving them semantically rich but spatially impoverished for 3D localization. We propose Voxel-enhanced Affordance detection (VoxAfford), which bypasses this bottleneck by injecting multi-scale geometric features from a frozen pre-trained 3D VQVAE encoder into the output tokens after generation. Each output token uses its affordance semantics as a query to retrieve relevant geometric patterns from its paired voxel scale via cross-attention, with a learned compatibility gate controlling the injection strength. The enhanced tokens are then aggregated into a spatially-aware affordance prompt through semantic-conditioned attention and propagated alongside per-point features to generate the final mask. Experiments on open-vocabulary affordance detection tasks show that VoxAfford achieves state-of-the-art performance with approximately an 8% improvement in mIoU, and real robot experiments confirm zero-shot transfer to novel objects.
format Preprint
id arxiv_https___arxiv_org_abs_2605_01365
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle VoxAfford: Multi-Scale Voxel-Token Fusion for Open-Vocabulary 3D Affordance Detection
Sun, Haowen
Zhang, Shaolong
Li, Mingyang
Ma, Chengzhong
Chen, Xinzhe
Cui, Qiongjie
Chen, Xingyu
Liu, Zeyang
Lan, Xuguang
Computer Vision and Pattern Recognition
Robotics
Open-vocabulary 3D affordance detection requires localizing interaction regions on point clouds given novel affordance descriptions. Recent methods extend multimodal large language models (MLLMs) with special output tokens that are decoded into segmentation masks. However, these tokens are produced through autoregressive generation, which models sequential dependencies rather than spatial neighborhood relations, leaving them semantically rich but spatially impoverished for 3D localization. We propose Voxel-enhanced Affordance detection (VoxAfford), which bypasses this bottleneck by injecting multi-scale geometric features from a frozen pre-trained 3D VQVAE encoder into the output tokens after generation. Each output token uses its affordance semantics as a query to retrieve relevant geometric patterns from its paired voxel scale via cross-attention, with a learned compatibility gate controlling the injection strength. The enhanced tokens are then aggregated into a spatially-aware affordance prompt through semantic-conditioned attention and propagated alongside per-point features to generate the final mask. Experiments on open-vocabulary affordance detection tasks show that VoxAfford achieves state-of-the-art performance with approximately an 8% improvement in mIoU, and real robot experiments confirm zero-shot transfer to novel objects.
title VoxAfford: Multi-Scale Voxel-Token Fusion for Open-Vocabulary 3D Affordance Detection
topic Computer Vision and Pattern Recognition
Robotics
url https://arxiv.org/abs/2605.01365