PLAF: Pixel-wise Language-Aligned Feature Extraction for Efficient 3D Scene Understanding

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Wen, Junjie, He, Junlin, Ma, Fei, Cui, Jinqiang
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition Robotics
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866915951481454592
author	Wen, Junjie He, Junlin Ma, Fei Cui, Jinqiang
author_facet	Wen, Junjie He, Junlin Ma, Fei Cui, Jinqiang
contents	Accurate open-vocabulary 3D scene understanding requires semantic representations that are both language-aligned and spatially precise at the pixel level, while remaining scalable when lifted to 3D space. However, existing representations struggle to jointly satisfy these requirements, and densely propagating pixel-wise semantics to 3D often results in substantial redundancy, leading to inefficient storage and querying in large-scale scenes. To address these challenges, we present \emph{PLAF}, a Pixel-wise Language-Aligned Feature extraction framework that enables dense and accurate semantic alignment in 2D without sacrificing open-vocabulary expressiveness. Building upon this representation, we further design an efficient semantic storage and querying scheme that significantly reduces redundancy across both 2D and 3D domains. Experimental results show that \emph{PLAF} provides a strong semantic foundation for accurate and efficient open-vocabulary 3D scene understanding. The codes are publicly available at https://github.com/RockWenJJ/PLAF.
format	Preprint
id	arxiv_https___arxiv_org_abs_2604_15770
institution	arXiv
publishDate	2026
record_format	arxiv
spellingShingle	PLAF: Pixel-wise Language-Aligned Feature Extraction for Efficient 3D Scene Understanding Wen, Junjie He, Junlin Ma, Fei Cui, Jinqiang Computer Vision and Pattern Recognition Robotics Accurate open-vocabulary 3D scene understanding requires semantic representations that are both language-aligned and spatially precise at the pixel level, while remaining scalable when lifted to 3D space. However, existing representations struggle to jointly satisfy these requirements, and densely propagating pixel-wise semantics to 3D often results in substantial redundancy, leading to inefficient storage and querying in large-scale scenes. To address these challenges, we present \emph{PLAF}, a Pixel-wise Language-Aligned Feature extraction framework that enables dense and accurate semantic alignment in 2D without sacrificing open-vocabulary expressiveness. Building upon this representation, we further design an efficient semantic storage and querying scheme that significantly reduces redundancy across both 2D and 3D domains. Experimental results show that \emph{PLAF} provides a strong semantic foundation for accurate and efficient open-vocabulary 3D scene understanding. The codes are publicly available at https://github.com/RockWenJJ/PLAF.
title	PLAF: Pixel-wise Language-Aligned Feature Extraction for Efficient 3D Scene Understanding
topic	Computer Vision and Pattern Recognition Robotics
url	https://arxiv.org/abs/2604.15770