Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
1 file | 56.18 MB
Add to list
  1. Search results
  2. Autoencoder-based image dimensionalit...

Autoencoder-based image dimensionality reduction methods

Nina Žižakić
(2022)
Author
Promoter
(UGent)
Organization
Project
Abstract
In this thesis, we study how images can be represented in a more compact way that still captures their most important features and preserves the similarities and dissimilarities between the images. These compact representations of images, also known as ‘image encodings’, allow us to identify similar images and image patches -- an operation very important for image processing and computer vision. By identifying similar image patches, we can perform operations such as image denoising, image inpainting, object tracking between frames in a video, and panorama image stitching. By identifying similar images, we can quickly retrieve images similar to a query image, for example, like in Google’s “search by image” feature. Throughout this thesis, we use a machine-learning--based method called autoencoder for learning these compact representations of images.
In dit proefschrift bestuderen we hoe beelden kunnen worden voorgesteld op een compactere manier die toch hun belangrijkste kenmerken vastlegt en de overeenkomsten en verschillen tussen de beelden behoudt. Deze compacte representaties van beelden, ook bekend als "beeldcoderingen", stellen ons in staat om gelijkaardige beelden en beeldpatronen te identificeren - een operatie die zeer belangrijk is voor beeldverwerking en computervisie. Door soortgelijke beeldpatronen te identificeren, kunnen we bewerkingen uitvoeren zoals beelddenoising, beeldinkleuring, objectopsporing tussen frames in een video en panorama-beeldverbinding. Door soortgelijke beelden te identificeren, kunnen we snel beelden terugvinden die lijken op een gezochte afbeelding, bijvoorbeeld zoals in de functie "zoeken op beeld" van Google. In dit proefschrift gebruiken we een op machinaal leren gebaseerde methode, de autoencoder, om deze compacte representaties van beelden te leren.

Downloads

  • Download
    NinaZizakic PhD final rev6 COMMENTS corrections3 more 30 FINAL F I N A L v2.pdf
    • full text (Published version)
    • |
    • open access
    • |
    • PDF
    • |
    • 56.18 MB

Citation

Please use this url to cite or link to this publication:

MLA
Žižakić, Nina. Autoencoder-Based Image Dimensionality Reduction Methods. Ghent University. Faculty of Engineering and Architecture, 2022.
APA
Žižakić, N. (2022). Autoencoder-based image dimensionality reduction methods. Ghent University. Faculty of Engineering and Architecture, Ghent, Belgium.
Chicago author-date
Žižakić, Nina. 2022. “Autoencoder-Based Image Dimensionality Reduction Methods.” Ghent, Belgium: Ghent University. Faculty of Engineering and Architecture.
Chicago author-date (all authors)
Žižakić, Nina. 2022. “Autoencoder-Based Image Dimensionality Reduction Methods.” Ghent, Belgium: Ghent University. Faculty of Engineering and Architecture.
Vancouver
1.
Žižakić N. Autoencoder-based image dimensionality reduction methods. [Ghent, Belgium]: Ghent University. Faculty of Engineering and Architecture; 2022.
IEEE
[1]
N. Žižakić, “Autoencoder-based image dimensionality reduction methods,” Ghent University. Faculty of Engineering and Architecture, Ghent, Belgium, 2022.
@phdthesis{01GJAPXMHGCPD9G2X30WR4MVH0,
  abstract     = {{In this thesis, we study how images can be represented in a more compact way that still captures their most important features and preserves the similarities and dissimilarities between the images. These compact representations of images, also known as ‘image encodings’, allow us to identify similar images and image patches -- an operation very important for image processing and computer vision. By identifying similar image patches, we can perform operations such as image denoising, image inpainting, object tracking between frames in a video, and panorama image stitching. By identifying similar images, we can quickly retrieve images similar to a query image, for example, like in Google’s “search by image” feature. Throughout this thesis, we use a machine-learning--based method called autoencoder for learning these compact representations of images.}},
  author       = {{Žižakić, Nina}},
  isbn         = {{9789463556422}},
  language     = {{eng}},
  pages        = {{XXII, 141}},
  publisher    = {{Ghent University. Faculty of Engineering and Architecture}},
  school       = {{Ghent University}},
  title        = {{Autoencoder-based image dimensionality reduction methods}},
  year         = {{2022}},
}