Unified model improves motion blur removal across cameras and scenes
UniDeblur: Unified Motion Deblurring via Fourier Convolution Learning
Computer Vision and Pattern Recognition
Summary
Photos often get blurry when the camera or objects move, and fixing this blur is tricky because it varies a lot depending on the camera and scene. The authors created a new approach that learns to reverse blur more universally by focusing on how blur works as a kind of image filter, instead of just trying to fix pixels directly. They trained their method on a large mix of different blurry images and showed it can effectively clear up many types of motion blur from various sources. This could help make clearer photos in many real-world situations.
What this means in practice
- •For mobile app developers: Integrate a robust blur removal model that handles diverse camera and motion blur scenarios into photo editing apps.$Commercial implications: Enables selling advanced photo enhancement features that work reliably across devices and conditions.
- •For digital forensics teams: Improve quality of motion-blurred images from various sources for better evidence analysis using a unified deblurring model.
Authors
Xintian Mao, Jiansheng Wang, Huaiyuan Wang, Qingli Li, Yan Wang
Abstract
Image deblurring is a fundamental task in computer vision, yet existing methods mainly specialize in isolated scenarios. Consequently, the deblurring performance of these models drops significantly across scenarios with different cameras and image contents. This motivates the development of a unified motion deblurring framework that can handle diverse blur patterns. To this end, we construct a multi-source dataset and develop a data pruning and quality control pipeline. Then, to restore blur images with various motion blur kernels across different scenarios, we propose a unified motion deblurring framework via Fourier Convolution Learning. Existing deblurring networks usually rely on pixel-level mappings and overlook the convolutional nature of blur. Thus, the network mostly fits the image content itself rather than learning the convolution process, making it difficult to handle multi-source blur images. Instead of relying on spatial mapping, we perform blur-adaptive global filter learning in Fourier space, which avoids the limited spatial support of kernel estimation. Extensive experiments demonstrate that our unified model effectively handles motion blurs across diverse sources and scenes, achieving competitive performance across both in-domain and OOD benchmarks. Code will be released at github.com/DeepMed-Lab-ECNU/Single-Image-Deblur.