Single image super-resolution using dictionary-based local regression

Sundaresh Ram, Jeffrey J Rodriguez

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Citations (Scopus)

Abstract

This paper presents a new method of producing a high-resolution image from a single low-resolution image without any external training image sets. We use a dictionary-based regression model for practical image super-resolution using local self-similar example patches within the image. Our method is inspired by the observation that image patches can be well represented as a sparse linear combination of elements from a chosen over-complete dictionary and that a patch in the high-resolution image have good matches around its corresponding location in the low-resolution image. A first-order approximation of a nonlinear mapping function, learned using the local self-similar example patches, is applied to the low-resolution image patches to obtain the corresponding high-resolution image patches. We show that the proposed algorithm provides improved accuracy compared to the existing single image super-resolution methods by running them on various input images that contain diverse textures, and that are contaminated by noise or other artifacts.

Original languageEnglish (US)
Title of host publicationProceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages121-124
Number of pages4
ISBN (Print)9781479940530
DOIs
StatePublished - 2014
Event2014 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2014 - San Diego, CA, United States
Duration: Apr 6 2014Apr 8 2014

Other

Other2014 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2014
CountryUnited States
CitySan Diego, CA
Period4/6/144/8/14

Fingerprint

Glossaries
Image resolution
Textures

Keywords

  • dictionary learning
  • Image restoration
  • image super-resolution
  • regression
  • sparse recovery

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Computer Science Applications

Cite this

Ram, S., & Rodriguez, J. J. (2014). Single image super-resolution using dictionary-based local regression. In Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation (pp. 121-124). [6806044] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SSIAI.2014.6806044

Single image super-resolution using dictionary-based local regression. / Ram, Sundaresh; Rodriguez, Jeffrey J.

Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation. Institute of Electrical and Electronics Engineers Inc., 2014. p. 121-124 6806044.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ram, S & Rodriguez, JJ 2014, Single image super-resolution using dictionary-based local regression. in Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation., 6806044, Institute of Electrical and Electronics Engineers Inc., pp. 121-124, 2014 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2014, San Diego, CA, United States, 4/6/14. https://doi.org/10.1109/SSIAI.2014.6806044
Ram S, Rodriguez JJ. Single image super-resolution using dictionary-based local regression. In Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation. Institute of Electrical and Electronics Engineers Inc. 2014. p. 121-124. 6806044 https://doi.org/10.1109/SSIAI.2014.6806044
Ram, Sundaresh ; Rodriguez, Jeffrey J. / Single image super-resolution using dictionary-based local regression. Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 121-124
@inproceedings{3987655f11ee445a99bb356150382ca4,
title = "Single image super-resolution using dictionary-based local regression",
abstract = "This paper presents a new method of producing a high-resolution image from a single low-resolution image without any external training image sets. We use a dictionary-based regression model for practical image super-resolution using local self-similar example patches within the image. Our method is inspired by the observation that image patches can be well represented as a sparse linear combination of elements from a chosen over-complete dictionary and that a patch in the high-resolution image have good matches around its corresponding location in the low-resolution image. A first-order approximation of a nonlinear mapping function, learned using the local self-similar example patches, is applied to the low-resolution image patches to obtain the corresponding high-resolution image patches. We show that the proposed algorithm provides improved accuracy compared to the existing single image super-resolution methods by running them on various input images that contain diverse textures, and that are contaminated by noise or other artifacts.",
keywords = "dictionary learning, Image restoration, image super-resolution, regression, sparse recovery",
author = "Sundaresh Ram and Rodriguez, {Jeffrey J}",
year = "2014",
doi = "10.1109/SSIAI.2014.6806044",
language = "English (US)",
isbn = "9781479940530",
pages = "121--124",
booktitle = "Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Single image super-resolution using dictionary-based local regression

AU - Ram, Sundaresh

AU - Rodriguez, Jeffrey J

PY - 2014

Y1 - 2014

N2 - This paper presents a new method of producing a high-resolution image from a single low-resolution image without any external training image sets. We use a dictionary-based regression model for practical image super-resolution using local self-similar example patches within the image. Our method is inspired by the observation that image patches can be well represented as a sparse linear combination of elements from a chosen over-complete dictionary and that a patch in the high-resolution image have good matches around its corresponding location in the low-resolution image. A first-order approximation of a nonlinear mapping function, learned using the local self-similar example patches, is applied to the low-resolution image patches to obtain the corresponding high-resolution image patches. We show that the proposed algorithm provides improved accuracy compared to the existing single image super-resolution methods by running them on various input images that contain diverse textures, and that are contaminated by noise or other artifacts.

AB - This paper presents a new method of producing a high-resolution image from a single low-resolution image without any external training image sets. We use a dictionary-based regression model for practical image super-resolution using local self-similar example patches within the image. Our method is inspired by the observation that image patches can be well represented as a sparse linear combination of elements from a chosen over-complete dictionary and that a patch in the high-resolution image have good matches around its corresponding location in the low-resolution image. A first-order approximation of a nonlinear mapping function, learned using the local self-similar example patches, is applied to the low-resolution image patches to obtain the corresponding high-resolution image patches. We show that the proposed algorithm provides improved accuracy compared to the existing single image super-resolution methods by running them on various input images that contain diverse textures, and that are contaminated by noise or other artifacts.

KW - dictionary learning

KW - Image restoration

KW - image super-resolution

KW - regression

KW - sparse recovery

UR - http://www.scopus.com/inward/record.url?scp=84902292695&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84902292695&partnerID=8YFLogxK

U2 - 10.1109/SSIAI.2014.6806044

DO - 10.1109/SSIAI.2014.6806044

M3 - Conference contribution

AN - SCOPUS:84902292695

SN - 9781479940530

SP - 121

EP - 124

BT - Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation

PB - Institute of Electrical and Electronics Engineers Inc.

ER -