Taming generative video models for zero-shot optical flow extraction

• URL: http://arxiv.org/abs/2507.09082v1
• Date: Fri, 11 Jul 2025 23:59:38 GMT
• Title: Taming generative video models for zero-shot optical flow extraction
• Authors: Seungwoo Kim, Khai Loong Aw, Klemen Kotar, Cristobal Eyzaguirre, Wanhee Lee, Yunong Liu, Jared Watrous, Stefan Stojanov, Juan Carlos Niebles, Jiajun Wu, Daniel L. K. Yamins,
• Abstract summary: Self-supervised video models trained only for future frame prediction can be prompted, without fine-tuning, to output flow.<n>Inspired by the Counterfactual World Model (CWM) paradigm, we extend this idea to generative video models.<n> KL-tracing is a novel test-time procedure that injects a localized perturbation into the first frame, rolls out the model one step, and computes the Kullback-Leibler divergence between perturbed and ungenerative predictive distributions.
• Score: 28.176290134216995
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Extracting optical flow from videos remains a core computer vision problem. Motivated by the success of large general-purpose models, we ask whether frozen self-supervised video models trained only for future frame prediction can be prompted, without fine-tuning, to output flow. Prior work reading out depth or illumination from video generators required fine-tuning, which is impractical for flow where labels are scarce and synthetic datasets suffer from a sim-to-real gap. Inspired by the Counterfactual World Model (CWM) paradigm, which can obtain point-wise correspondences by injecting a small tracer perturbation into a next-frame predictor and tracking its propagation, we extend this idea to generative video models. We explore several popular architectures and find that successful zero-shot flow extraction in this manner is aided by three model properties: (1) distributional prediction of future frames (avoiding blurry or noisy outputs); (2) factorized latents that treat each spatio-temporal patch independently; and (3) random-access decoding that can condition on any subset of future pixels. These properties are uniquely present in the recent Local Random Access Sequence (LRAS) architecture. Building on LRAS, we propose KL-tracing: a novel test-time procedure that injects a localized perturbation into the first frame, rolls out the model one step, and computes the Kullback-Leibler divergence between perturbed and unperturbed predictive distributions. Without any flow-specific fine-tuning, our method outperforms state-of-the-art models on real-world TAP-Vid DAVIS dataset (16.6% relative improvement for endpoint error) and synthetic TAP-Vid Kubric (4.7% relative improvement). Our results indicate that counterfactual prompting of controllable generative video models is a scalable and effective alternative to supervised or photometric-loss approaches for high-quality flow.

Related papers

• Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion [70.4360995984905]
  We introduce Self Forcing, a novel training paradigm for autoregressive video diffusion models.<n>It addresses the longstanding issue of exposure bias, where models trained on ground-truth context must generate sequences conditioned on their own imperfect outputs.
  arXiv  Detail & Related papers  (2025-06-09T17:59:55Z)
• Solving Inverse Problems with FLAIR [59.02385492199431]
  Flow-based latent generative models are able to generate images with remarkable quality, even enabling text-to-image generation.<n>We present FLAIR, a novel training free variational framework that leverages flow-based generative models as a prior for inverse problems.<n>Results on standard imaging benchmarks demonstrate that FLAIR consistently outperforms existing diffusion- and flow-based methods in terms of reconstruction quality and sample diversity.
  arXiv  Detail & Related papers  (2025-06-03T09:29:47Z)
• Autoregressive Video Generation without Vector Quantization [90.87907377618747]
  We reformulate the video generation problem as a non-quantized autoregressive modeling of temporal frame-by-frame prediction.<n>With the proposed approach, we train a novel video autoregressive model without vector quantization, termed NOVA.<n>Our results demonstrate that NOVA surpasses prior autoregressive video models in data efficiency, inference speed, visual fidelity, and video fluency, even with a much smaller model capacity.
  arXiv  Detail & Related papers  (2024-12-18T18:59:53Z)
• From Slow Bidirectional to Fast Autoregressive Video Diffusion Models [52.32078428442281]
  Current video diffusion models achieve impressive generation quality but struggle in interactive applications due to bidirectional attention dependencies.<n>We address this limitation by adapting a pretrained bidirectional diffusion transformer to an autoregressive transformer that generates frames on-the-fly.<n>Our model achieves a total score of 84.27 on the VBench-Long benchmark, surpassing all previous video generation models.
  arXiv  Detail & Related papers  (2024-12-10T18:59:50Z)
• Efficient Continuous Video Flow Model for Video Prediction [43.16308241800144]
  Multi-step prediction models, such as diffusion and rectified flow models, exhibit higher latency in sampling new frames compared to single-step methods.<n>We propose a novel approach to modeling the multi-step process, aimed at alleviating latency constraints and facilitating the adaptation of such processes for video prediction tasks.
  arXiv  Detail & Related papers  (2024-12-07T12:11:25Z)
• SIGMA:Sinkhorn-Guided Masked Video Modeling [69.31715194419091]
  Sinkhorn-guided Masked Video Modelling ( SIGMA) is a novel video pretraining method. We distribute features of space-time tubes evenly across a limited number of learnable clusters. Experimental results on ten datasets validate the effectiveness of SIGMA in learning more performant, temporally-aware, and robust video representations.
  arXiv  Detail & Related papers  (2024-07-22T08:04:09Z)
• Efficient Video Prediction via Sparsely Conditioned Flow Matching [24.32740918613266]
  We introduce a novel generative model for video prediction based on latent flow matching. We call our model Random frame conditioned flow Integration for VidEo pRediction, or, in short, RIVER.
  arXiv  Detail & Related papers  (2022-11-26T14:18:50Z)
• Unsupervised Flow-Aligned Sequence-to-Sequence Learning for Video Restoration [85.3323211054274]
  How to properly model the inter-frame relation within the video sequence is an important but unsolved challenge for video restoration (VR) In this work, we propose an unsupervised flow-aligned sequence-to-sequence model (S2SVR) to address this problem. S2SVR shows superior performance in multiple VR tasks, including video deblurring, video super-resolution, and compressed video quality enhancement.
  arXiv  Detail & Related papers  (2022-05-20T14:14:48Z)
• A Log-likelihood Regularized KL Divergence for Video Prediction with A 3D Convolutional Variational Recurrent Network [17.91970304953206]
  We introduce a new variational model that extends the recurrent network in two ways for the task of frame prediction. First, we introduce 3D convolutions inside all modules including the recurrent model for future prediction frame, inputting sequence and outputting video frames at each timestep. Second, we enhance the latent loss predictions of the variational model by introducing a maximum likelihood estimate in addition to the KL that is commonly used in variational models.
  arXiv  Detail & Related papers  (2020-12-11T05:05:31Z)

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.