ICG-MVSNet: Learning Intra-view and Cross-view Relationships for Guidance in Multi-View Stereo

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Hu, Yuxi, Zhang, Jun, Zhang, Zhe, Weilharter, Rafael, Rao, Yuchen, Chen, Kuangyi, Yuan, Runze, Fraundorfer, Friedrich
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910896664608768
author Hu, Yuxi
Zhang, Jun
Zhang, Zhe
Weilharter, Rafael
Rao, Yuchen
Chen, Kuangyi
Yuan, Runze
Fraundorfer, Friedrich
author_facet Hu, Yuxi
Zhang, Jun
Zhang, Zhe
Weilharter, Rafael
Rao, Yuchen
Chen, Kuangyi
Yuan, Runze
Fraundorfer, Friedrich
contents Multi-view Stereo (MVS) aims to estimate depth and reconstruct 3D point clouds from a series of overlapping images. Recent learning-based MVS frameworks overlook the geometric information embedded in features and correlations, leading to weak cost matching. In this paper, we propose ICG-MVSNet, which explicitly integrates intra-view and cross-view relationships for depth estimation. Specifically, we develop an intra-view feature fusion module that leverages the feature coordinate correlations within a single image to enhance robust cost matching. Additionally, we introduce a lightweight cross-view aggregation module that efficiently utilizes the contextual information from volume correlations to guide regularization. Our method is evaluated on the DTU dataset and Tanks and Temples benchmark, consistently achieving competitive performance against state-of-the-art works, while requiring lower computational resources.
format Preprint
id arxiv_https___arxiv_org_abs_2503_21525
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ICG-MVSNet: Learning Intra-view and Cross-view Relationships for Guidance in Multi-View Stereo
Hu, Yuxi
Zhang, Jun
Zhang, Zhe
Weilharter, Rafael
Rao, Yuchen
Chen, Kuangyi
Yuan, Runze
Fraundorfer, Friedrich
Computer Vision and Pattern Recognition
Multi-view Stereo (MVS) aims to estimate depth and reconstruct 3D point clouds from a series of overlapping images. Recent learning-based MVS frameworks overlook the geometric information embedded in features and correlations, leading to weak cost matching. In this paper, we propose ICG-MVSNet, which explicitly integrates intra-view and cross-view relationships for depth estimation. Specifically, we develop an intra-view feature fusion module that leverages the feature coordinate correlations within a single image to enhance robust cost matching. Additionally, we introduce a lightweight cross-view aggregation module that efficiently utilizes the contextual information from volume correlations to guide regularization. Our method is evaluated on the DTU dataset and Tanks and Temples benchmark, consistently achieving competitive performance against state-of-the-art works, while requiring lower computational resources.
title ICG-MVSNet: Learning Intra-view and Cross-view Relationships for Guidance in Multi-View Stereo
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.21525