MegaFlow: Zero-Shot Large Displacement Optical Flow

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhang, Dingxi, Wang, Fangjinhua, Pollefeys, Marc, Xu, Haofei
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908915889864704
author Zhang, Dingxi
Wang, Fangjinhua
Pollefeys, Marc
Xu, Haofei
author_facet Zhang, Dingxi
Wang, Fangjinhua
Pollefeys, Marc
Xu, Haofei
contents Accurate estimation of large displacement optical flow remains a critical challenge. Existing methods typically rely on iterative local search or/and domain-specific fine-tuning, which severely limits their performance in large displacement and zero-shot generalization scenarios. To overcome this, we introduce MegaFlow, a simple yet powerful model for zero-shot large displacement optical flow. Rather than relying on highly complex, task-specific architectural designs, MegaFlow adapts powerful pre-trained vision priors to produce temporally consistent motion fields. In particular, we formulate flow estimation as a global matching problem by leveraging pre-trained global Vision Transformer features, which naturally capture large displacements. This is followed by a few lightweight iterative refinements to further improve the sub-pixel accuracy. Extensive experiments demonstrate that MegaFlow achieves state-of-the-art zero-shot performance across multiple optical flow benchmarks. Moreover, our model also delivers highly competitive zero-shot performance on long-range point tracking benchmarks, demonstrating its robust transferability and suggesting a unified paradigm for generalizable motion estimation. Our project page is at: https://kristen-z.github.io/projects/megaflow.
format Preprint
id arxiv_https___arxiv_org_abs_2603_25739
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MegaFlow: Zero-Shot Large Displacement Optical Flow
Zhang, Dingxi
Wang, Fangjinhua
Pollefeys, Marc
Xu, Haofei
Computer Vision and Pattern Recognition
Accurate estimation of large displacement optical flow remains a critical challenge. Existing methods typically rely on iterative local search or/and domain-specific fine-tuning, which severely limits their performance in large displacement and zero-shot generalization scenarios. To overcome this, we introduce MegaFlow, a simple yet powerful model for zero-shot large displacement optical flow. Rather than relying on highly complex, task-specific architectural designs, MegaFlow adapts powerful pre-trained vision priors to produce temporally consistent motion fields. In particular, we formulate flow estimation as a global matching problem by leveraging pre-trained global Vision Transformer features, which naturally capture large displacements. This is followed by a few lightweight iterative refinements to further improve the sub-pixel accuracy. Extensive experiments demonstrate that MegaFlow achieves state-of-the-art zero-shot performance across multiple optical flow benchmarks. Moreover, our model also delivers highly competitive zero-shot performance on long-range point tracking benchmarks, demonstrating its robust transferability and suggesting a unified paradigm for generalizable motion estimation. Our project page is at: https://kristen-z.github.io/projects/megaflow.
title MegaFlow: Zero-Shot Large Displacement Optical Flow
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.25739