GS-RealBlur: A Flexible Data Acquisition Framework for Real-World Image Deblurring
Computer Vision and Pattern Recognition
Summary
The authors created a new way to get realistic blurry and sharp image pairs for training computer programs that fix blurry pictures. They used a handheld camera to take blurry photos and a gimbal to quickly take many sharp photos of the same scene. Then, they built a 3D model to match each blurry photo with a sharp image from the same viewpoint. They also developed a method to improve how these images are aligned. Their dataset helps train models that work better on real blurry images than previous datasets. They plan to share their data and code with others.
Authors
Mingyang Chen, Zhilu Zhang, Honglei Xu, Renlong Wu, Xiaohe Wu, Wangmeng Zuo
Abstract
High-quality, large-scale paired data is essential for training learning-based image deblurring models. However, synthetic blurry images generally lack realism, while real-world captured images require complex and inflexible camera systems. In this work, we propose GS-RealBlur, a data acquisition framework for real-world image deblurring, achieving both blur realism and acquisition flexibility. Specifically, we use a handheld camera to capture blurry images, and deploy a gimbal to densely capture sharp images of the same scene. We reconstruct the 3D representation of sharp images and calibrate the camera pose of each blurry frame within this 3D. The image rendered from this 3D according to the pose serves as the sharp counterpart. To better align the rendered image with the blurry image, we introduce a Blur-aware Pose Refinement (BPR) module that refines the pose using appearance consistency and centroid alignment constraints. Leveraging GS-RealBlur, we construct a high-quality and diverse dataset. Extensive experiments demonstrate that a deblurring model trained on our dataset achieves superior generalization performance across various real-world deblurring benchmarks, consistently outperforming models trained on existing synthetic and real-world datasets. The code and dataset will be made publicly available.