Presentation 2014-12-19
Image Regularization with Morphological Gradient Priors
Makoto NAKASHIZUKA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, a regularization method based on morphological gradients is proposed for image recovery problem. In our approach, the image prior of the regularization is defined as the sum of the morphological gradients. The morphological gradient is obtained as the difference between the dilation and erosion, which are respectively obtained by the max and min filtering of the image. The property of the morphological gradient depends on the structuring elements of the erosion and dilation. There is a possibility to specify an appropriate structuring element for recovery of a specific image. For minimization of the image recovery, we apply the subgradient method to the objective function that includes the morphological gradient. In order to show some property of the image recovery with the morphological gradients, denoising examples with various structuring elements are demonstrated. Comparisons of the proposed prior with total variation (TV) prior are also shown.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Morphology / morphological gradient / regularization / image recovery / convex optimization / sub-gradient method
Paper # SIS2014-81
Date of Issue

Conference Information
Committee SIS
Conference Date 2014/12/11(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Smart Info-Media Systems (SIS)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Image Regularization with Morphological Gradient Priors
Sub Title (in English)
Keyword(1) Morphology
Keyword(2) morphological gradient
Keyword(3) regularization
Keyword(4) image recovery
Keyword(5) convex optimization
Keyword(6) sub-gradient method
1st Author's Name Makoto NAKASHIZUKA
1st Author's Affiliation Faculty of Engineering, Chiba Institute of Technology()
Date 2014-12-19
Paper # SIS2014-81
Volume (vol) vol.114
Number (no) 370
Page pp.pp.-
#Pages 6
Date of Issue