MVSMamba: Multi-View Stereo with State Space Model

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Jiang, Jianfei, Liu, Qiankun, Liu, Hongyuan, Yu, Haochen, Wang, Liyong, Chen, Jiansheng, Ma, Huimin
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917055500910592
author Jiang, Jianfei
Liu, Qiankun
Liu, Hongyuan
Yu, Haochen
Wang, Liyong
Chen, Jiansheng
Ma, Huimin
author_facet Jiang, Jianfei
Liu, Qiankun
Liu, Hongyuan
Yu, Haochen
Wang, Liyong
Chen, Jiansheng
Ma, Huimin
contents Robust feature representations are essential for learning-based Multi-View Stereo (MVS), which relies on accurate feature matching. Recent MVS methods leverage Transformers to capture long-range dependencies based on local features extracted by conventional feature pyramid networks. However, the quadratic complexity of Transformer-based MVS methods poses challenges to balance performance and efficiency. Motivated by the global modeling capability and linear complexity of the Mamba architecture, we propose MVSMamba, the first Mamba-based MVS network. MVSMamba enables efficient global feature aggregation with minimal computational overhead. To fully exploit Mamba's potential in MVS, we propose a Dynamic Mamba module (DM-module) based on a novel reference-centered dynamic scanning strategy, which enables: (1) Efficient intra- and inter-view feature interaction from the reference to source views, (2) Omnidirectional multi-view feature representations, and (3) Multi-scale global feature aggregation. Extensive experimental results demonstrate MVSMamba outperforms state-of-the-art MVS methods on the DTU dataset and the Tanks-and-Temples benchmark with both superior performance and efficiency. The source code is available at https://github.com/JianfeiJ/MVSMamba.
format Preprint
id arxiv_https___arxiv_org_abs_2511_01315
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MVSMamba: Multi-View Stereo with State Space Model
Jiang, Jianfei
Liu, Qiankun
Liu, Hongyuan
Yu, Haochen
Wang, Liyong
Chen, Jiansheng
Ma, Huimin
Computer Vision and Pattern Recognition
Robust feature representations are essential for learning-based Multi-View Stereo (MVS), which relies on accurate feature matching. Recent MVS methods leverage Transformers to capture long-range dependencies based on local features extracted by conventional feature pyramid networks. However, the quadratic complexity of Transformer-based MVS methods poses challenges to balance performance and efficiency. Motivated by the global modeling capability and linear complexity of the Mamba architecture, we propose MVSMamba, the first Mamba-based MVS network. MVSMamba enables efficient global feature aggregation with minimal computational overhead. To fully exploit Mamba's potential in MVS, we propose a Dynamic Mamba module (DM-module) based on a novel reference-centered dynamic scanning strategy, which enables: (1) Efficient intra- and inter-view feature interaction from the reference to source views, (2) Omnidirectional multi-view feature representations, and (3) Multi-scale global feature aggregation. Extensive experimental results demonstrate MVSMamba outperforms state-of-the-art MVS methods on the DTU dataset and the Tanks-and-Temples benchmark with both superior performance and efficiency. The source code is available at https://github.com/JianfeiJ/MVSMamba.
title MVSMamba: Multi-View Stereo with State Space Model
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.01315