StableMotion: Repurposing Diffusion-Based Image Priors for Motion Estimation
This work addresses motion estimation problems in computer vision, offering a novel application of diffusion models with significant speed improvements, though it is incremental in leveraging existing models.
The authors tackled single-image-based image rectification tasks, such as Stitched Image Rectangling and Rolling Shutter Correction, by repurposing pretrained diffusion models for motion estimation, achieving state-of-the-art performance and a 200x speedup over previous diffusion-based methods.
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.