Physics-Grounded Motion Forecasting via Equation Discovery for Trajectory-Guided Image-to-Video Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Feng, Tao, Zhao, Xianbing, Chen, Zhenhua, Wong, Tien Tsin, Rezatofighi, Hamid, Haffari, Gholamreza, Qu, Lizhen
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912473065455616
author Feng, Tao
Zhao, Xianbing
Chen, Zhenhua
Wong, Tien Tsin
Rezatofighi, Hamid
Haffari, Gholamreza
Qu, Lizhen
author_facet Feng, Tao
Zhao, Xianbing
Chen, Zhenhua
Wong, Tien Tsin
Rezatofighi, Hamid
Haffari, Gholamreza
Qu, Lizhen
contents Recent advances in diffusion-based and autoregressive video generation models have achieved remarkable visual realism. However, these models typically lack accurate physical alignment, failing to replicate real-world dynamics in object motion. This limitation arises primarily from their reliance on learned statistical correlations rather than capturing mechanisms adhering to physical laws. To address this issue, we introduce a novel framework that integrates symbolic regression (SR) and trajectory-guided image-to-video (I2V) models for physics-grounded video forecasting. Our approach extracts motion trajectories from input videos, uses a retrieval-based pre-training mechanism to enhance symbolic regression, and discovers equations of motion to forecast physically accurate future trajectories. These trajectories then guide video generation without requiring fine-tuning of existing models. Evaluated on scenarios in Classical Mechanics, including spring-mass, pendulums, and projectile motions, our method successfully recovers ground-truth analytical equations and improves the physical alignment of generated videos over baseline methods.
format Preprint
id arxiv_https___arxiv_org_abs_2507_06830
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Physics-Grounded Motion Forecasting via Equation Discovery for Trajectory-Guided Image-to-Video Generation
Feng, Tao
Zhao, Xianbing
Chen, Zhenhua
Wong, Tien Tsin
Rezatofighi, Hamid
Haffari, Gholamreza
Qu, Lizhen
Computer Vision and Pattern Recognition
Artificial Intelligence
Recent advances in diffusion-based and autoregressive video generation models have achieved remarkable visual realism. However, these models typically lack accurate physical alignment, failing to replicate real-world dynamics in object motion. This limitation arises primarily from their reliance on learned statistical correlations rather than capturing mechanisms adhering to physical laws. To address this issue, we introduce a novel framework that integrates symbolic regression (SR) and trajectory-guided image-to-video (I2V) models for physics-grounded video forecasting. Our approach extracts motion trajectories from input videos, uses a retrieval-based pre-training mechanism to enhance symbolic regression, and discovers equations of motion to forecast physically accurate future trajectories. These trajectories then guide video generation without requiring fine-tuning of existing models. Evaluated on scenarios in Classical Mechanics, including spring-mass, pendulums, and projectile motions, our method successfully recovers ground-truth analytical equations and improves the physical alignment of generated videos over baseline methods.
title Physics-Grounded Motion Forecasting via Equation Discovery for Trajectory-Guided Image-to-Video Generation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2507.06830