Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution

Jakub Adler; Jolanta Kawulok; Michal Kawulok

doi:10.1007/978-3-031-21967-2_27

Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution

Department of Algorithmics and Software (RAU/RAU5)

Silesian University of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Citation (Scopus)

Abstract

Super-resolution reconstruction is a common term for a variety of techniques aimed at enhancing spatial resolution either from a single image or from multiple images presenting the same scene. While single-image super-resolution has been intensively explored with many advancements proposed attributed to the use of deep learning, multi-image reconstruction remains a much less explored field. The first solutions based on convolutional neural networks were proposed recently for super-resolving multiple Proba-V satellite images, but they have not been validated for enhancing natural images so far. Also, their sensitiveness to the characteristics of the input data, including their mutual similarity and image acquisition conditions, has not been explored in depth. In this paper, we address this research gap to better understand how to select and prepare the input data for reconstruction. We expect that the reported conclusions will help in elaborating more efficient super-resolution frameworks that could be deployed in practical applications.

Original language	English
Title of host publication	Intelligent Information and Database Systems - 14th Asian Conference, ACIIDS 2022, Proceedings
Editors	Ngoc Thanh Nguyen, Bogdan Trawiński, Ngoc Thanh Nguyen, Tien Khoa Tran, Ualsher Tukayev, Tzung-Pei Hong, Edward Szczerbicki
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	329-342
Number of pages	14
ISBN (Print)	9783031219665
DOIs	https://doi.org/10.1007/978-3-031-21967-2_27
Publication status	Published - 2022
Event	14th Asian Conference on Intelligent Information and Database Systems , ACIIDS 2022 - Ho Chi Minh City, Viet Nam Duration: 28 Nov 2022 → 30 Nov 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13758 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th Asian Conference on Intelligent Information and Database Systems , ACIIDS 2022
Country/Territory	Viet Nam
City	Ho Chi Minh City
Period	28/11/22 → 30/11/22

Keywords

Convolutional neural networks
Data selection
Deep learning
Multi-image super-resolution
Super-resolution

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-21967-2_27

Cite this

Adler, J., Kawulok, J., & Kawulok, M. (2022). Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution. In N. T. Nguyen, B. Trawiński, N. T. Nguyen, T. K. Tran, U. Tukayev, T.-P. Hong, & E. Szczerbicki (Eds.), Intelligent Information and Database Systems - 14th Asian Conference, ACIIDS 2022, Proceedings (pp. 329-342). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13758 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-21967-2_27

Adler, Jakub ; Kawulok, Jolanta ; Kawulok, Michal. / Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution. Intelligent Information and Database Systems - 14th Asian Conference, ACIIDS 2022, Proceedings. editor / Ngoc Thanh Nguyen ; Bogdan Trawiński ; Ngoc Thanh Nguyen ; Tien Khoa Tran ; Ualsher Tukayev ; Tzung-Pei Hong ; Edward Szczerbicki. Springer Science and Business Media Deutschland GmbH, 2022. pp. 329-342 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{e5c2e86156c14ed99fc66b9ed6277b11,

title = "Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution",

abstract = "Super-resolution reconstruction is a common term for a variety of techniques aimed at enhancing spatial resolution either from a single image or from multiple images presenting the same scene. While single-image super-resolution has been intensively explored with many advancements proposed attributed to the use of deep learning, multi-image reconstruction remains a much less explored field. The first solutions based on convolutional neural networks were proposed recently for super-resolving multiple Proba-V satellite images, but they have not been validated for enhancing natural images so far. Also, their sensitiveness to the characteristics of the input data, including their mutual similarity and image acquisition conditions, has not been explored in depth. In this paper, we address this research gap to better understand how to select and prepare the input data for reconstruction. We expect that the reported conclusions will help in elaborating more efficient super-resolution frameworks that could be deployed in practical applications.",

keywords = "Convolutional neural networks, Data selection, Deep learning, Multi-image super-resolution, Super-resolution",

author = "Jakub Adler and Jolanta Kawulok and Michal Kawulok",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 14th Asian Conference on Intelligent Information and Database Systems , ACIIDS 2022 ; Conference date: 28-11-2022 Through 30-11-2022",

year = "2022",

doi = "10.1007/978-3-031-21967-2\_27",

language = "English",

isbn = "9783031219665",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "329--342",

editor = "Nguyen, \{Ngoc Thanh\} and Bogdan Trawi{\'n}ski and Nguyen, \{Ngoc Thanh\} and Tran, \{Tien Khoa\} and Ualsher Tukayev and Tzung-Pei Hong and Edward Szczerbicki",

booktitle = "Intelligent Information and Database Systems - 14th Asian Conference, ACIIDS 2022, Proceedings",

address = "Germany",

}

Adler, J, Kawulok, J & Kawulok, M 2022, Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution. in NT Nguyen, B Trawiński, NT Nguyen, TK Tran, U Tukayev, T-P Hong & E Szczerbicki (eds), Intelligent Information and Database Systems - 14th Asian Conference, ACIIDS 2022, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13758 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 329-342, 14th Asian Conference on Intelligent Information and Database Systems , ACIIDS 2022, Ho Chi Minh City, Viet Nam, 28/11/22. https://doi.org/10.1007/978-3-031-21967-2_27

Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution. / Adler, Jakub; Kawulok, Jolanta ; Kawulok, Michal.
Intelligent Information and Database Systems - 14th Asian Conference, ACIIDS 2022, Proceedings. ed. / Ngoc Thanh Nguyen; Bogdan Trawiński; Ngoc Thanh Nguyen; Tien Khoa Tran; Ualsher Tukayev; Tzung-Pei Hong; Edward Szczerbicki. Springer Science and Business Media Deutschland GmbH, 2022. p. 329-342 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13758 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution

AU - Adler, Jakub

AU - Kawulok, Jolanta

AU - Kawulok, Michal

PY - 2022

Y1 - 2022

N2 - Super-resolution reconstruction is a common term for a variety of techniques aimed at enhancing spatial resolution either from a single image or from multiple images presenting the same scene. While single-image super-resolution has been intensively explored with many advancements proposed attributed to the use of deep learning, multi-image reconstruction remains a much less explored field. The first solutions based on convolutional neural networks were proposed recently for super-resolving multiple Proba-V satellite images, but they have not been validated for enhancing natural images so far. Also, their sensitiveness to the characteristics of the input data, including their mutual similarity and image acquisition conditions, has not been explored in depth. In this paper, we address this research gap to better understand how to select and prepare the input data for reconstruction. We expect that the reported conclusions will help in elaborating more efficient super-resolution frameworks that could be deployed in practical applications.

AB - Super-resolution reconstruction is a common term for a variety of techniques aimed at enhancing spatial resolution either from a single image or from multiple images presenting the same scene. While single-image super-resolution has been intensively explored with many advancements proposed attributed to the use of deep learning, multi-image reconstruction remains a much less explored field. The first solutions based on convolutional neural networks were proposed recently for super-resolving multiple Proba-V satellite images, but they have not been validated for enhancing natural images so far. Also, their sensitiveness to the characteristics of the input data, including their mutual similarity and image acquisition conditions, has not been explored in depth. In this paper, we address this research gap to better understand how to select and prepare the input data for reconstruction. We expect that the reported conclusions will help in elaborating more efficient super-resolution frameworks that could be deployed in practical applications.

KW - Convolutional neural networks

KW - Data selection

KW - Deep learning

KW - Multi-image super-resolution

KW - Super-resolution

UR - https://www.scopus.com/pages/publications/85145265317

U2 - 10.1007/978-3-031-21967-2_27

DO - 10.1007/978-3-031-21967-2_27

M3 - Conference contribution

AN - SCOPUS:85145265317

SN - 9783031219665

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 329

EP - 342

BT - Intelligent Information and Database Systems - 14th Asian Conference, ACIIDS 2022, Proceedings

A2 - Nguyen, Ngoc Thanh

A2 - Trawiński, Bogdan

A2 - Nguyen, Ngoc Thanh

A2 - Tran, Tien Khoa

A2 - Tukayev, Ualsher

A2 - Hong, Tzung-Pei

A2 - Szczerbicki, Edward

PB - Springer Science and Business Media Deutschland GmbH

T2 - 14th Asian Conference on Intelligent Information and Database Systems , ACIIDS 2022

Y2 - 28 November 2022 through 30 November 2022

ER -

Adler J, Kawulok J , Kawulok M. Toward Understanding the Impact of Input Data for Multi-Image Super-Resolution. In Nguyen NT, Trawiński B, Nguyen NT, Tran TK, Tukayev U, Hong TP, Szczerbicki E, editors, Intelligent Information and Database Systems - 14th Asian Conference, ACIIDS 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 329-342. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-21967-2_27