Publication detail

Image demosaicing using Deep Image Prior

BALUŠÍK, P.

Original Title

Image demosaicing using Deep Image Prior

Type

article in a collection out of WoS and Scopus

Language

English

Original Abstract

The paper focuses on the problem of image demosaicing using the deep image prior. The deep image prior (DIP) is an uncommon concept that uses a generative neural network which, however, utilizes only the degraded image as the input for training. A novel method for image demosaicing is proposed, based on DIP, and it is compared with common demosaicing methods. In terms of the objective PSNR and SSIM values, the proposed method proved to be comparable with a widely used Malvar’s demosaicing method. Nevertheless, subjectively, DIP produces demosaiced images comparable with the superior Menon’s algorithm. Unfortunately, the proposed method turned out to be computationally immensely challenging.

Keywords

demosaicing, debayerization, color filter array, deep image prior

Authors

BALUŠÍK, P.

Released

25. 4. 2023

Publisher

Brno University of Technology, Faculty of Electrical Engineering and Communication

Location

Brno

ISBN

978-80-214-6154-3

Book

Proceedings II of the 29th Conference STUDENT EEICT 2023 Selected papers

Edition

1

ISBN

2788-1334

Periodical

Proceedings II of the Conference STUDENT EEICT

State

Czech Republic

Pages from

17

Pages to

20

Pages count

4

URL

BibTex

@inproceedings{BUT184280,
  author="Peter {Balušík}",
  title="Image demosaicing using Deep Image Prior",
  booktitle="Proceedings II of the 29th Conference STUDENT EEICT 2023 Selected papers",
  year="2023",
  series="1",
  journal="Proceedings II of the Conference STUDENT EEICT",
  pages="17--20",
  publisher="Brno University of Technology, Faculty of Electrical Engineering and Communication",
  address="Brno",
  doi="10.13164/eeict.2023.17",
  isbn="978-80-214-6154-3",
  issn="2788-1334",
  url="https://www.eeict.cz/eeict_download/archiv/sborniky/EEICT_2023_sbornik_2_v2.pdf"
}