Learning to restore multiple image degradations simultaneously

Zhang, Le; Bronik, Kevin; Papież, Bartłomiej W.

doi:10.1016/j.patcog.2022.109250

Cited by 5 publications

(1 citation statement)

References 74 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous works develop various techniques to improve deblurring performance [1,2,3]. They effectively restore a clear image in some simple conditions, e.g., uniform blur, or specific scenarios (human face or text image).…”

Section: Introductionmentioning

confidence: 99%

Take a Prior from Other Tasks for Severe Blur Removal

Wang¹,

Xue²,

Zhu³

et al. 2023

Preprint

View full text Add to dashboard Cite

Recovering clear structures from severely blurry inputs is a challenging problem due to the large movements between the camera and the scene. Although some works apply segmentation maps on human face images for deblurring, they cannot handle natural scenes because objects and degradation are more complex, and inaccurate segmentation maps lead to a loss of details. For general scene deblurring, the feature space of the blurry image and corresponding sharp image under the high-level vision task is closer, which inspires us to rely on other tasks (e.g. classification) to learn a comprehensive prior in severe blur removal cases. We propose a cross-level feature learning strategy based on knowledge distillation to learn the priors, which include global contexts and sharp local structures for recovering potential details.In addition, we propose a semantic prior embedding layer with multi-level aggregation and semantic attention transformation to integrate the priors effectively. We introduce the proposed priors to various models, including the UNet and other mainstream deblurring baselines, leading to better per-

show abstract

Section: Introductionmentioning

confidence: 99%