Restricted Boltzmann machine approach to couple dictionary training for image super-resolution

Junbin Gao; Yi Guo; Ming Yin

doi:10.1109/ICIP.2013.6738103

Restricted Boltzmann machine approach to couple dictionary training for image super-resolution

Junbin Gao, Yi Guo, Ming Yin

Western Sydney University

Research output: Chapter in Book / Conference Paper › Conference Paper › peer-review

14 Citations (Scopus)

Abstract

Image super-resolution means forming high-resolution images from low-resolution images. In this paper, we develop a new approach based on the deep Restricted Boltzmann Machines (RBM) for image super-resolution. The RBM architecture has ability of learning a set of visual patterns, called dictionary elements from a set of training images. The learned dictionary will be then used to synthesize high resolution images. We test the proposed algorithm on both benchmark and natural images, comparing with several other techniques. The visual quality of the results has also been assessed by both human evaluation and quantitative measurement.

Original language	English
Title of host publication	Proceedings ICIP 2013: 2013 IEEE International Conference on Image Processing, 15-18 September 2013, Melbourne, Victoria, Australia
Publisher	IEEE
Pages	499-503
Number of pages	5
ISBN (Print)	9781479923410
DOIs	https://doi.org/10.1109/ICIP.2013.6738103
Publication status	Published - 2013
Event	International Conference on Image Processing - Duration: 15 Sept 2013 → …

Conference

Conference	International Conference on Image Processing
Period	15/09/13 → …

Keywords

algorithms
high resolution imaging
resolution (optics)

Access to Document

10.1109/ICIP.2013.6738103

Cite this

@inproceedings{c51361bf7d304b09b9549bfbb41ce6ea,

title = "Restricted Boltzmann machine approach to couple dictionary training for image super-resolution",

abstract = "Image super-resolution means forming high-resolution images from low-resolution images. In this paper, we develop a new approach based on the deep Restricted Boltzmann Machines (RBM) for image super-resolution. The RBM architecture has ability of learning a set of visual patterns, called dictionary elements from a set of training images. The learned dictionary will be then used to synthesize high resolution images. We test the proposed algorithm on both benchmark and natural images, comparing with several other techniques. The visual quality of the results has also been assessed by both human evaluation and quantitative measurement.",

keywords = "algorithms, high resolution imaging, resolution (optics)",

author = "Junbin Gao and Yi Guo and Ming Yin",

year = "2013",

doi = "10.1109/ICIP.2013.6738103",

language = "English",

isbn = "9781479923410",

pages = "499--503",

booktitle = "Proceedings ICIP 2013: 2013 IEEE International Conference on Image Processing, 15-18 September 2013, Melbourne, Victoria, Australia",

publisher = "IEEE",

note = "International Conference on Image Processing ; Conference date: 15-09-2013",

}

Gao, J, Guo, Y & Yin, M 2013, Restricted Boltzmann machine approach to couple dictionary training for image super-resolution. in Proceedings ICIP 2013: 2013 IEEE International Conference on Image Processing, 15-18 September 2013, Melbourne, Victoria, Australia. IEEE, pp. 499-503, International Conference on Image Processing, 15/09/13. https://doi.org/10.1109/ICIP.2013.6738103

Restricted Boltzmann machine approach to couple dictionary training for image super-resolution. / Gao, Junbin; Guo, Yi; Yin, Ming.
Proceedings ICIP 2013: 2013 IEEE International Conference on Image Processing, 15-18 September 2013, Melbourne, Victoria, Australia. IEEE, 2013. p. 499-503.

Research output: Chapter in Book / Conference Paper › Conference Paper › peer-review

TY - GEN

T1 - Restricted Boltzmann machine approach to couple dictionary training for image super-resolution

AU - Gao, Junbin

AU - Guo, Yi

AU - Yin, Ming

PY - 2013

Y1 - 2013

N2 - Image super-resolution means forming high-resolution images from low-resolution images. In this paper, we develop a new approach based on the deep Restricted Boltzmann Machines (RBM) for image super-resolution. The RBM architecture has ability of learning a set of visual patterns, called dictionary elements from a set of training images. The learned dictionary will be then used to synthesize high resolution images. We test the proposed algorithm on both benchmark and natural images, comparing with several other techniques. The visual quality of the results has also been assessed by both human evaluation and quantitative measurement.

AB - Image super-resolution means forming high-resolution images from low-resolution images. In this paper, we develop a new approach based on the deep Restricted Boltzmann Machines (RBM) for image super-resolution. The RBM architecture has ability of learning a set of visual patterns, called dictionary elements from a set of training images. The learned dictionary will be then used to synthesize high resolution images. We test the proposed algorithm on both benchmark and natural images, comparing with several other techniques. The visual quality of the results has also been assessed by both human evaluation and quantitative measurement.

KW - algorithms

KW - high resolution imaging

KW - resolution (optics)

UR - http://handle.uws.edu.au:8081/1959.7/uws:35235

UR - https://www2.securecms.com/ICIP2013/

U2 - 10.1109/ICIP.2013.6738103

DO - 10.1109/ICIP.2013.6738103

M3 - Conference Paper

SN - 9781479923410

SP - 499

EP - 503

BT - Proceedings ICIP 2013: 2013 IEEE International Conference on Image Processing, 15-18 September 2013, Melbourne, Victoria, Australia

PB - IEEE

T2 - International Conference on Image Processing

Y2 - 15 September 2013

ER -

Restricted Boltzmann machine approach to couple dictionary training for image super-resolution

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this