Physics in 2-Steps: Locking Motion Priors Before Visual Refinement Erases Them
Summary
PhaseLock is a training-free framework that preserves motion priors from early-step inference to improve physical consistency in image-to-video diffusion models, achieving 6.2 point improvement with minimal overhead.
View Cached Full Text
Cached at: 06/08/26, 11:15 AM
Paper page - Physics in 2-Steps: Locking Motion Priors Before Visual Refinement Erases Them
Source: https://huggingface.co/papers/2606.06361
Abstract
PhaseLock is a training-free framework that improves physical consistency in image-to-video diffusion models by preserving motion priors from early-step inference throughout the denoising process.
Image-to-Video diffusion modelsleverage input images to generate visually stunning content, yet frequently produce motion that violates physical laws. We reveal a surprising finding: a 2-step generation often exhibits betterphysical consistencythan a 50-step output from the same model. Throughspectral analysis, we trace this tophase erosionduringdenoising; the phase degrades significantly (dropping by approx 18% from step 2 to step 50), whereas the magnitude remains relatively stable. Building on this insight, we propose PhaseLock, a training-free framework that preserves the validmotion priorsfrom few-step inference throughout thedenoisingtrajectory. Rather than relying on full-step inference forphysical consistency, PhaseLock extracts a motion prior from just 2 steps and enforces it onto high-fidelity generation viaLatent Delta Guidance. Our approach effectively mitigates phase degradation, improvingphysical consistencyby an average of 6.2 points across diverse models while largely maintainingvisual fidelity, with negligible overhead (1.06times time, 1.02times memory) and reduced reliance on expensive external guidance methods (sim5times time).
View arXiv pageView PDFProject pageGitHub2Add to collection
Get this paper in your agent:
hf papers read 2606\.06361
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2606.06361 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2606.06361 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2606.06361 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
PhyMotion: Structured 3D Motion Reward for Physics-Grounded Human Video Generation
PhyMotion proposes a physics-grounded reward system that evaluates kinematic plausibility, contact consistency, and dynamic feasibility of human motion in generated videos, achieving stronger correlation with human judgment and improving motion realism in RL-based post-training.
ResilPhase: Plug-and-Play Phase Mapping and Noise-Resilient Macro-Trajectory Extrapolation for Diffusion Acceleration
ResilPhase is a training-free acceleration framework for diffusion models that reformulates accelerated inference as stable macro-trajectory extrapolation in ODE space, using derivative-free barycentric Lagrange extrapolation and bounded phase mapping to achieve state-of-the-art fidelity under high acceleration ratios.
Steady-Forcing: Balancing Spatial Persistence and Motion Continuity in Long-Horizon Nature Video Diffusion
Steady-Forcing proposes a memory and training framework to balance spatial stability and motion continuity in long-horizon nature video generation, improving background consistency while sustaining fluid dynamics over multi-minute rollouts.
Diagnosing and Repairing Shape-Prior Shortcuts in Long-Range Single-Shot Fringe Projection Profilometry
This paper diagnoses and repairs shape-prior shortcuts in learning-based long-range single-shot fringe projection profilometry, using mechanistic interpretability and conformal uncertainty quantification. The proposed PhiCalNet architecture achieves a 3.3x reduction in object MAE by replacing depth regression with wrapped-phase output and a differentiable calibration layer.
PhysisForcing: Physics Reinforced World Simulator for Robotic Manipulation
PhysisForcing is a training framework that enhances embodied video generation for robotic manipulation by enforcing physical consistency through pixel-level trajectory alignment and semantic-level relational alignment losses in a DiT-based architecture, achieving notable improvements on benchmarks.