MegaFlow: Zero-Shot Large Displacement Optical Flow

• URL: http://arxiv.org/abs/2603.25739v1
• Date: Thu, 26 Mar 2026 17:59:51 GMT
• Title: MegaFlow: Zero-Shot Large Displacement Optical Flow
• Abstract summary: MegaFlow is a simple yet powerful model for zero-shot large displacement optical flow.<n>It adapts powerful pre-trained vision priors to produce temporally consistent motion fields.<n>It achieves state-of-the-art zero-shot performance across multiple optical flow benchmarks.
• Score: 54.89185614376759
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• 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.

Related papers

• FlowIt: Global Matching for Optical Flow with Confidence-Guided Refinement [44.07406912385784]
  FlowIt is a novel architecture for optical flow estimation.<n>It is designed to robustly handle large pixel displacements.<n>It achieves state-of-the-art results on the competitive Sintel and KITTI benchmarks.
  arXiv  Detail & Related papers  (2026-03-30T17:58:12Z)
• FlowFeat: Pixel-Dense Embedding of Motion Profiles [51.933932255458906]
  FlowFeat is a high-resolution and multi-task feature representation.<n>Our work takes a step forward towards reliable and versatile dense image representations.
  arXiv  Detail & Related papers  (2025-11-10T23:39:53Z)
• FlowIE: Efficient Image Enhancement via Rectified Flow [71.6345505427213]
  FlowIE is a flow-based framework that estimates straight-line paths from an elementary distribution to high-quality images. Our contributions are rigorously validated through comprehensive experiments on synthetic and real-world datasets.
  arXiv  Detail & Related papers  (2024-06-01T17:29:29Z)
• Sparse Global Matching for Video Frame Interpolation with Large Motion [20.49084881829404]
  Large motion poses a critical challenge in Video Frame Interpolation (VFI) task. Existing methods are often constrained by limited receptive fields, resulting in sub-optimal performance when handling scenarios with large motion. We introduce a new pipeline for VFI, which can effectively integrate global-level information to alleviate issues associated with large motion.
  arXiv  Detail & Related papers  (2024-04-10T11:06:29Z)
• GAFlow: Incorporating Gaussian Attention into Optical Flow [62.646389181507764]
  We push Gaussian Attention (GA) into the optical flow models to accentuate local properties during representation learning. We introduce a novel Gaussian-Constrained Layer (GCL) which can be easily plugged into existing Transformer blocks. For reliable motion analysis, we provide a new Gaussian-Guided Attention Module (GGAM)
  arXiv  Detail & Related papers  (2023-09-28T07:46:01Z)
• AccFlow: Backward Accumulation for Long-Range Optical Flow [70.4251045372285]
  This paper proposes a novel recurrent framework called AccFlow for long-range optical flow estimation. We demonstrate the superiority of backward accumulation over conventional forward accumulation. Experiments validate the effectiveness of AccFlow in handling long-range optical flow estimation.
  arXiv  Detail & Related papers  (2023-08-25T01:51:26Z)
• Towards Anytime Optical Flow Estimation with Event Cameras [37.892840680923584]
  Event cameras respond to changes in log-brightness at the millisecond level, making them ideal for optical flow estimation.<n>Existing datasets from event cameras provide only low frame rate ground truth for optical flow.<n>We propose EVA-Flow, an EVent-based Anytime Flow estimation network to produce high-frame-rate event optical flow.
  arXiv  Detail & Related papers  (2023-07-11T06:15:12Z)
• AnyFlow: Arbitrary Scale Optical Flow with Implicit Neural Representation [17.501820140334328]
  We introduce AnyFlow, a robust network that estimates accurate flow from images of various resolutions. We establish a new state-of-the-art performance of cross-dataset generalization on the KITTI dataset.
  arXiv  Detail & Related papers  (2023-03-29T07:03:51Z)
• GMFlow: Learning Optical Flow via Global Matching [124.57850500778277]
  We propose a GMFlow framework for learning optical flow estimation. It consists of three main components: a customized Transformer for feature enhancement, a correlation and softmax layer for global feature matching, and a self-attention layer for flow propagation. Our new framework outperforms 32-iteration RAFT's performance on the challenging Sintel benchmark.
  arXiv  Detail & Related papers  (2021-11-26T18:59:56Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.