Estimation of the regularisation parameter in Huber-MRF for image resolution enhancement

The Huber Markov Random Field (H-MRF) has been proposed for image resolution enhancement as a preferable alternative to Gaussian Random Markov Fields (G-MRF) for its ability to preserve discontinuities in the image. However, its performance relies on a good choice of a regularisation parameter. Whil...

Full description

Saved in:
Bibliographic Details
Main Authors: S., Ali Pitchay,, A., Kabán,
Format: Conference Paper
Language:en_US
Published: 2015
Subjects:
Online Access:http://ddms.usim.edu.my/handle/123456789/9197
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.usim-9197
record_format dspace
spelling my.usim-91972015-08-25T04:38:10Z Estimation of the regularisation parameter in Huber-MRF for image resolution enhancement S., Ali Pitchay, A., Kabán, Huber prior Hyper-parameter Optimisation Super-resolution The Huber Markov Random Field (H-MRF) has been proposed for image resolution enhancement as a preferable alternative to Gaussian Random Markov Fields (G-MRF) for its ability to preserve discontinuities in the image. However, its performance relies on a good choice of a regularisation parameter. While automating this choice has been successfully tackled for G-MRF, the more sophisticated form of H-MRF makes this problem less straightforward. In this paper we develop an approximate solution to this problem, by upper-bounding the partition function of the H-MRF. We demonstrate the working and flexibility of our approach in image super-resolution experiments. © 2013 Springer-Verlag. 2015-08-25T04:38:10Z 2015-08-25T04:38:10Z 2013-01-01 Conference Paper 9783-6424-1277-6 3029-743 http://ddms.usim.edu.my/handle/123456789/9197 en_US
institution Universiti Sains Islam Malaysia
building USIM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universit Sains Islam i Malaysia
content_source USIM Institutional Repository
url_provider http://ddms.usim.edu.my/
language en_US
topic Huber prior
Hyper-parameter Optimisation
Super-resolution
spellingShingle Huber prior
Hyper-parameter Optimisation
Super-resolution
S., Ali Pitchay,
A., Kabán,
Estimation of the regularisation parameter in Huber-MRF for image resolution enhancement
description The Huber Markov Random Field (H-MRF) has been proposed for image resolution enhancement as a preferable alternative to Gaussian Random Markov Fields (G-MRF) for its ability to preserve discontinuities in the image. However, its performance relies on a good choice of a regularisation parameter. While automating this choice has been successfully tackled for G-MRF, the more sophisticated form of H-MRF makes this problem less straightforward. In this paper we develop an approximate solution to this problem, by upper-bounding the partition function of the H-MRF. We demonstrate the working and flexibility of our approach in image super-resolution experiments. © 2013 Springer-Verlag.
format Conference Paper
author S., Ali Pitchay,
A., Kabán,
author_facet S., Ali Pitchay,
A., Kabán,
author_sort S., Ali Pitchay,
title Estimation of the regularisation parameter in Huber-MRF for image resolution enhancement
title_short Estimation of the regularisation parameter in Huber-MRF for image resolution enhancement
title_full Estimation of the regularisation parameter in Huber-MRF for image resolution enhancement
title_fullStr Estimation of the regularisation parameter in Huber-MRF for image resolution enhancement
title_full_unstemmed Estimation of the regularisation parameter in Huber-MRF for image resolution enhancement
title_sort estimation of the regularisation parameter in huber-mrf for image resolution enhancement
publishDate 2015
url http://ddms.usim.edu.my/handle/123456789/9197
_version_ 1645152560540024832
score 13.209306