Non-uniform motion deblurring with blurry component divided guidance
作者：
Highlights：
• A two-branch component divided guidance model is proposed for non-uniform deblurring.
• An attention module adaptively divides the blur into large and small blurry regions.
• An orientation-based fusion module to merge the two branches’ de-blurred features.
• Considering the guidance delivers a remarkable performance.
• The parameters and computation of the model are smaller than other methods.
摘要
•A two-branch component divided guidance model is proposed for non-uniform deblurring.•An attention module adaptively divides the blur into large and small blurry regions.•An orientation-based fusion module to merge the two branches’ de-blurred features.•Considering the guidance delivers a remarkable performance.•The parameters and computation of the model are smaller than other methods.
论文关键词：Non-uiniform deblurring,Component divided,Attention mechanism
论文评审过程：Received 9 January 2021, Revised 30 April 2021, Accepted 26 May 2021, Available online 10 June 2021, Version of Record 27 June 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2021.108082
