Abstract:Deep learning (DL) based methods for motion deblurring, taking advantage of large-scale datasets and sophisticated network structures, have reported promising results. However, two challenges still remain: existing methods usually perform well on synthetic datasets but cannot deal with complex real-world blur, and in addition, over- and under-estimation of the blur will result in restored images that remain blurred and even introduce unwanted distortion. We propose a motion deblurring framework that includes a Blur Space Disentangled Network (BSDNet) and a Hierarchical Scale-recurrent Deblurring Network (HSDNet) to address these issues. Specifically, we train an image blurring model to facilitate learning a better image deblurring model. Firstly, BSDNet learns how to separate the blur features from blurry images, which is adaptable for blur transferring, dataset augmentation, and ultimately directing the deblurring model. Secondly, to gradually recover sharp information in a coarse-to-fine manner, HSDNet makes full use of the blur features acquired by BSDNet as a priori and breaks down the non-uniform deblurring task into various subtasks. Moreover, the motion blur dataset created by BSDNet also bridges the gap between training images and actual blur. Extensive experiments on real-world blur datasets demonstrate that our method works effectively on complex scenarios, resulting in the best performance that significantly outperforms many state-of-the-art approaches.

World from Blur.

Planar Motion Deblurring from Dual Images Using Transparency

Image Deblurring Utilizing Inertial Sensors and a Short-Long-Short Exposure Strategy.

Exposure Trajectory Recovery from Motion Blur

Recovering Continuous Scene Dynamics from A Single Blurry Image with Events

Image Deblurring With Image Blurring

Rethinking Blur Synthesis for Deep Real-World Image Deblurring

Motion Deblurring in the Wild

Event-Assisted Blurriness Representation Learning for Blurry Image Unfolding

Single Image Deblurring and Camera Motion Estimation with Depth Map

Perceptual Variousness Motion Deblurring with Light Global Context Refinement

Learning Event-Based Motion Deblurring

Event-Assisted Recurrent Network for Arbitrary-Temporal-Scale Blurry Image Unfolding

Every Moment Matters: Detail-Aware Networks to Bring a Blurry Image Alive

Motion Blur Decomposition with Cross-shutter Guidance

Self-Supervised Linear Motion Deblurring

Blind Nonuniform Deblur under Projection Motion Path

Blur-Attention: A boosting mechanism for non-uniform blurred image restoration

DeblurGS: Gaussian Splatting for Camera Motion Blur

Learning Blind Motion Deblurring

DyBluRF: Dynamic Neural Radiance Fields from Blurry Monocular Video