MVL-Loc: Leveraging Vision-Language Model for Generalizable Multi-Scene Camera Relocalization

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Xiao, Zhendong, Wei, Wu, Ji, Shujie, Yang, Shan, Chen, Changhao
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911041989902336
author Xiao, Zhendong
Wei, Wu
Ji, Shujie
Yang, Shan
Chen, Changhao
author_facet Xiao, Zhendong
Wei, Wu
Ji, Shujie
Yang, Shan
Chen, Changhao
contents Camera relocalization, a cornerstone capability of modern computer vision, accurately determines a camera's position and orientation (6-DoF) from images and is essential for applications in augmented reality (AR), mixed reality (MR), autonomous driving, delivery drones, and robotic navigation. Unlike traditional deep learning-based methods that regress camera pose from images in a single scene, which often lack generalization and robustness in diverse environments, we propose MVL-Loc, a novel end-to-end multi-scene 6-DoF camera relocalization framework. MVL-Loc leverages pretrained world knowledge from vision-language models (VLMs) and incorporates multimodal data to generalize across both indoor and outdoor settings. Furthermore, natural language is employed as a directive tool to guide the multi-scene learning process, facilitating semantic understanding of complex scenes and capturing spatial relationships among objects. Extensive experiments on the 7Scenes and Cambridge Landmarks datasets demonstrate MVL-Loc's robustness and state-of-the-art performance in real-world multi-scene camera relocalization, with improved accuracy in both positional and orientational estimates.
format Preprint
id arxiv_https___arxiv_org_abs_2507_04509
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MVL-Loc: Leveraging Vision-Language Model for Generalizable Multi-Scene Camera Relocalization
Xiao, Zhendong
Wei, Wu
Ji, Shujie
Yang, Shan
Chen, Changhao
Computer Vision and Pattern Recognition
Artificial Intelligence
Camera relocalization, a cornerstone capability of modern computer vision, accurately determines a camera's position and orientation (6-DoF) from images and is essential for applications in augmented reality (AR), mixed reality (MR), autonomous driving, delivery drones, and robotic navigation. Unlike traditional deep learning-based methods that regress camera pose from images in a single scene, which often lack generalization and robustness in diverse environments, we propose MVL-Loc, a novel end-to-end multi-scene 6-DoF camera relocalization framework. MVL-Loc leverages pretrained world knowledge from vision-language models (VLMs) and incorporates multimodal data to generalize across both indoor and outdoor settings. Furthermore, natural language is employed as a directive tool to guide the multi-scene learning process, facilitating semantic understanding of complex scenes and capturing spatial relationships among objects. Extensive experiments on the 7Scenes and Cambridge Landmarks datasets demonstrate MVL-Loc's robustness and state-of-the-art performance in real-world multi-scene camera relocalization, with improved accuracy in both positional and orientational estimates.
title MVL-Loc: Leveraging Vision-Language Model for Generalizable Multi-Scene Camera Relocalization
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2507.04509