Planning with Sketch-Guided Verification for Physics-Aware Video Generation

TL;DR

This paper introduces SketchVerify, a planning framework for physics-aware video generation that uses sketch-based verification to produce more physically plausible and instruction-consistent motions efficiently before full video synthesis.

Contribution

The paper presents a training-free, test-time sketch-verification method that improves motion planning quality in video generation by ranking and refining candidate trajectories with a lightweight verification process.

Findings

Significantly improves motion quality and physical realism in generated videos.

Achieves better long-term consistency compared to baseline methods.

Enhances efficiency by avoiding expensive diffusion-based synthesis during planning.

Abstract

Recent video generation approaches increasingly rely on planning intermediate control signals such as object trajectories to improve temporal coherence and motion fidelity. However, these methods mostly employ single-shot plans that are typically limited to simple motions, or iterative refinement which requires multiple calls to the video generator, incuring high computational cost. To overcome these limitations, we propose SketchVerify, a training-free, sketch-verification-based planning framework that improves motion planning quality with more dynamically coherent trajectories (i.e., physically plausible and instruction-consistent motions) prior to full video generation by introducing a test-time sampling and verification loop. Given a prompt and a reference image, our method predicts multiple candidate motion plans and ranks them using a vision-language verifier that jointly evaluates semantic alignment with the instruction and physical plausibility. To efficiently score candidate motion plans, we render each trajectory as a lightweight video sketch by compositing objects over a static background, which bypasses the need for expensive, repeated diffusion-based synthesis while achieving comparable performance. We iteratively refine the motion plan until a satisfactory one is identified, which is then passed to the trajectory-conditioned generator for final synthesis. Experiments on WorldModelBench and PhyWorldBench demonstrate that our method significantly improves motion quality, physical realism, and long-term consistency compared to competitive baselines while being substantially more efficient. Our ablation study further shows that scaling up the number of trajectory candidates consistently enhances overall performance.