Defocus map estimation from a single image using improved likelihood feature and edge-based basis
作者：
Highlights：
• Edge-based and region-based methods are combined via RTF with linear basis.
• Orthogonal gradients with Gabor filter subsets can obtain more accurate likelihood.
• First K highest local maximums are effective to serve as the input feature of RTF.
• Proposed method outperforms state-of-the-art DME methods.
• Proposed method is readily applied to image deblurring and defocus blur detection.
摘要
•Edge-based and region-based methods are combined via RTF with linear basis.•Orthogonal gradients with Gabor filter subsets can obtain more accurate likelihood.•First K highest local maximums are effective to serve as the input feature of RTF.•Proposed method outperforms state-of-the-art DME methods.•Proposed method is readily applied to image deblurring and defocus blur detection.
论文关键词：Defocus map estimation,Regression tree fields,Localized 2D frequency analysis
论文评审过程：Received 12 August 2018, Revised 18 March 2020, Accepted 3 June 2020, Available online 6 June 2020, Version of Record 12 June 2020.
论文官网地址：https://doi.org/10.1016/j.patcog.2020.107485
