Learning non-negativity constrained variation for image denoising and deblurring

Tengda Wei; Linshan Wang; Ping Lin; Jialing Chen; Yangfan Wang; Haiyong Zheng

doi:10.4208/nmtma.2017.m1653

Learning non-negativity constrained variation for image denoising and deblurring

Tengda Wei
, Linshan Wang
, Ping Lin
, Jialing Chen
, Yangfan Wang (Lead / Corresponding author)
, Haiyong Zheng

Mathematics

Research output: Contribution to journal › Article › peer-review

4 Citations (Scopus)

463 Downloads (Pure)

Abstract

This paper presents a heuristic Learning-based Non-Negativity Constrained Variation (L-NNCV) aiming to search the coefficients of variational model automatically and make the variation adapt different images and problems by supervised-learning strategy. The model includes two terms: a problem-based term that is derived from the prior knowledge, and an image-driven regularization which is learned by some training samples. The model can be solved by classical o-constraint method. Experimental results show that: the experimental effectiveness of each term in the regularization accords with the corresponding theoretical proof; the proposed method outperforms other PDE-based methods on image denoising and deblurring.

Original language	English
Pages (from-to)	852-871
Number of pages	20
Journal	Numerical Mathematics: Theory, Methods and Applications
Volume	10
Issue number	4
Early online date	12 Sept 2017
DOIs	https://doi.org/10.4208/nmtma.2017.m1653
Publication status	Published - 1 Nov 2017

Keywords

Constraint
Image restoration
Learning idea
O-constraint method
TV-based model

ASJC Scopus subject areas

Modelling and Simulation
Control and Optimization
Computational Mathematics
Applied Mathematics

Access to Document

10.4208/nmtma.2017.m1653

Author Accepted Manuscript
This article has been published in a revised form in Numerical Mathematics: Theory, Methods and Applications [http://dx.doi.org/10.4208/nmtma.2017.m1653]. This version is free to view and download for private research and study only. Not for re-distribution, re-sale or use in derivative works. © Global-Science Press 2017.
Accepted author manuscript, 6.2 MB

Cite this

@article{d3b96ffa0d684375a58bd3d37c8351c4,

title = "Learning non-negativity constrained variation for image denoising and deblurring",

abstract = "This paper presents a heuristic Learning-based Non-Negativity Constrained Variation (L-NNCV) aiming to search the coefficients of variational model automatically and make the variation adapt different images and problems by supervised-learning strategy. The model includes two terms: a problem-based term that is derived from the prior knowledge, and an image-driven regularization which is learned by some training samples. The model can be solved by classical o-constraint method. Experimental results show that: the experimental effectiveness of each term in the regularization accords with the corresponding theoretical proof; the proposed method outperforms other PDE-based methods on image denoising and deblurring.",

keywords = "Constraint, Image restoration, Learning idea, O-constraint method, TV-based model",

author = "Tengda Wei and Linshan Wang and Ping Lin and Jialing Chen and Yangfan Wang and Haiyong Zheng",

year = "2017",

month = nov,

day = "1",

doi = "10.4208/nmtma.2017.m1653",

language = "English",

volume = "10",

pages = "852--871",

journal = "Numerical Mathematics: Theory, Methods and Applications",

issn = "1004-8979",

publisher = "Global Science Press",

number = "4",

}

TY - JOUR

T1 - Learning non-negativity constrained variation for image denoising and deblurring

AU - Wei, Tengda

AU - Wang, Linshan

AU - Lin, Ping

AU - Chen, Jialing

AU - Wang, Yangfan

AU - Zheng, Haiyong

PY - 2017/11/1

Y1 - 2017/11/1

N2 - This paper presents a heuristic Learning-based Non-Negativity Constrained Variation (L-NNCV) aiming to search the coefficients of variational model automatically and make the variation adapt different images and problems by supervised-learning strategy. The model includes two terms: a problem-based term that is derived from the prior knowledge, and an image-driven regularization which is learned by some training samples. The model can be solved by classical o-constraint method. Experimental results show that: the experimental effectiveness of each term in the regularization accords with the corresponding theoretical proof; the proposed method outperforms other PDE-based methods on image denoising and deblurring.

AB - This paper presents a heuristic Learning-based Non-Negativity Constrained Variation (L-NNCV) aiming to search the coefficients of variational model automatically and make the variation adapt different images and problems by supervised-learning strategy. The model includes two terms: a problem-based term that is derived from the prior knowledge, and an image-driven regularization which is learned by some training samples. The model can be solved by classical o-constraint method. Experimental results show that: the experimental effectiveness of each term in the regularization accords with the corresponding theoretical proof; the proposed method outperforms other PDE-based methods on image denoising and deblurring.

KW - Constraint

KW - Image restoration

KW - Learning idea

KW - O-constraint method

KW - TV-based model

U2 - 10.4208/nmtma.2017.m1653

DO - 10.4208/nmtma.2017.m1653

M3 - Article

AN - SCOPUS:85029916012

SN - 1004-8979

VL - 10

SP - 852

EP - 871

JO - Numerical Mathematics: Theory, Methods and Applications

JF - Numerical Mathematics: Theory, Methods and Applications

IS - 4

ER -

Learning non-negativity constrained variation for image denoising and deblurring

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this