MegaFlow: Zero-Shot Large Displacement Optical Flow

TL;DR

MegaFlow introduces a zero-shot large displacement optical flow model that leverages pre-trained vision transformers and global matching, achieving state-of-the-art performance without task-specific training.

Contribution

The paper presents MegaFlow, a novel approach that uses pre-trained vision priors and global matching for zero-shot large displacement optical flow estimation.

Findings

Achieves state-of-the-art zero-shot performance on optical flow benchmarks.

Demonstrates robust transferability to long-range point tracking.

Provides a unified paradigm for generalizable motion estimation.

Abstract

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.