We consider the problem of 3D shape recovery from ultra-fast motion-blurred images. While 3D reconstruction from static images has been extensively studied, recovering geometry from extreme motion-blurred images remains challenging. Such scenarios frequently occur in both natural and industrial settings, such as fast-moving objects in sports (e.g., balls) or rotating machinery, where rapid motion distorts object appearance and makes traditional 3D reconstruction techniques like Multi-View Stereo (MVS) ineffective.
In this paper, we propose a novel inverse rendering approach for shape recovery from ultra-fast motion-blurred images. While conventional rendering techniques typically synthesize blur by averaging across multiple frames, we identify a major computational bottleneck in the repeated computation of barycentric weights. To address this, we propose a fast barycentric coordinate solver, which significantly reduces computational overhead and achieves a speedup of up to 4.57×, enabling efficient and photorealistic simulation of high-speed motion. Crucially, our method is fully differentiable, allowing gradients to propagate from rendered images to the underlying 3D shape, thereby facilitating shape recovery through inverse rendering.
We validate our approach on two representative motion types: rapid translation and rotation. Experimental results demonstrate that our method enables efficient and realistic modeling of ultra-fast moving objects in the forward simulation. Moreover, it successfully recovers 3D shapes from 2D imagery of objects undergoing extreme translational and rotational motion, advancing the boundaries of vision-based 3D reconstruction.
Our method successfully recovers 3D shapes from motion-blurred images of rotating objects. Below shows the comparison between ground truth, input blurred images, and our reconstructed results across different objects.
Our method also recovers both geometry and appearance from translational motion blur. The results demonstrate high-quality reconstruction compared to ground truth.
Our rendered images and gradients exhibit high similarity to SoftRas across various motion cases and sample numbers, validating the correctness of our fast solver.
We validate our method on real captured motion-blurred images using a 100Hz high-speed camera and controlled motion rig.
Below shows more examples of real-world motion blur recovery, compared with the state-of-the-art method.
@inproceedings{yu2026recovering,
title={Recovering 3D Shapes from Ultra-Fast Motion-Blurred Images},
author={Yu, Fei and Guo, Shudan and Xin, Shiqing and Wang, Beibei and Zhao, Haisen and Chen, Wenzheng},
booktitle={International Conference on 3D Vision (3DV)},
year={2026}
}