StableMotion: Repurposing Diffusion-Based Image Priors for Motion Estimation

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Wang, Ziyi, Li, Haipeng, Sui, Lin, Zhou, Tianhao, Jiang, Hai, Nie, Lang, Liu, Shuaicheng
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915282478432256
author Wang, Ziyi
Li, Haipeng
Sui, Lin
Zhou, Tianhao
Jiang, Hai
Nie, Lang
Liu, Shuaicheng
author_facet Wang, Ziyi
Li, Haipeng
Sui, Lin
Zhou, Tianhao
Jiang, Hai
Nie, Lang
Liu, Shuaicheng
contents We present StableMotion, a novel framework leverages knowledge (geometry and content priors) from pretrained large-scale image diffusion models to perform motion estimation, solving single-image-based image rectification tasks such as Stitched Image Rectangling (SIR) and Rolling Shutter Correction (RSC). Specifically, StableMotion framework takes text-to-image Stable Diffusion (SD) models as backbone and repurposes it into an image-to-motion estimator. To mitigate inconsistent output produced by diffusion models, we propose Adaptive Ensemble Strategy (AES) that consolidates multiple outputs into a cohesive, high-fidelity result. Additionally, we present the concept of Sampling Steps Disaster (SSD), the counterintuitive scenario where increasing the number of sampling steps can lead to poorer outcomes, which enables our framework to achieve one-step inference. StableMotion is verified on two image rectification tasks and delivers state-of-the-art performance in both, as well as showing strong generalizability. Supported by SSD, StableMotion offers a speedup of 200 times compared to previous diffusion model-based methods.
format Preprint
id arxiv_https___arxiv_org_abs_2505_06668
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle StableMotion: Repurposing Diffusion-Based Image Priors for Motion Estimation
Wang, Ziyi
Li, Haipeng
Sui, Lin
Zhou, Tianhao
Jiang, Hai
Nie, Lang
Liu, Shuaicheng
Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
We present StableMotion, a novel framework leverages knowledge (geometry and content priors) from pretrained large-scale image diffusion models to perform motion estimation, solving single-image-based image rectification tasks such as Stitched Image Rectangling (SIR) and Rolling Shutter Correction (RSC). Specifically, StableMotion framework takes text-to-image Stable Diffusion (SD) models as backbone and repurposes it into an image-to-motion estimator. To mitigate inconsistent output produced by diffusion models, we propose Adaptive Ensemble Strategy (AES) that consolidates multiple outputs into a cohesive, high-fidelity result. Additionally, we present the concept of Sampling Steps Disaster (SSD), the counterintuitive scenario where increasing the number of sampling steps can lead to poorer outcomes, which enables our framework to achieve one-step inference. StableMotion is verified on two image rectification tasks and delivers state-of-the-art performance in both, as well as showing strong generalizability. Supported by SSD, StableMotion offers a speedup of 200 times compared to previous diffusion model-based methods.
title StableMotion: Repurposing Diffusion-Based Image Priors for Motion Estimation
topic Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
url https://arxiv.org/abs/2505.06668