Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach

<p dir="ltr">Researchers gravitate towards Generative Adversarial Networks (GAN) to create artificial images. However, GANs suffer from convergence issues, mode collapse, and overall complexity in balancing the Nash Equilibrium. Images generated are often distorted, rendering them us...

Full description

Saved in:
Bibliographic Details
Main Author: Ashhadul Islam (16869981) (author)
Other Authors: Samir Brahim Belhaouari (9427347) (author)
Published: 2023
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1864513527572594688
author Ashhadul Islam (16869981)
author2 Samir Brahim Belhaouari (9427347)
author2_role author
author_facet Ashhadul Islam (16869981)
Samir Brahim Belhaouari (9427347)
author_role author
dc.creator.none.fl_str_mv Ashhadul Islam (16869981)
Samir Brahim Belhaouari (9427347)
dc.date.none.fl_str_mv 2023-03-20T03:00:00Z
dc.identifier.none.fl_str_mv 10.1109/access.2023.3259236
dc.relation.none.fl_str_mv https://figshare.com/articles/journal_contribution/Fast_and_Efficient_Image_Generation_Using_Variational_Autoencoders_and_K-Nearest_Neighbor_OveRsampling_Approach/25204208
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Engineering
Electrical engineering
Electronics, sensors and digital hardware
Materials engineering
Generators
Feature extraction
Generative adversarial networks
Decoding
Training
Nash equilibrium
Image synthesis
Face recognition
variational autoencoders
Image reconstruction
artificial image creation
dc.title.none.fl_str_mv Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach
dc.type.none.fl_str_mv Text
Journal contribution
info:eu-repo/semantics/publishedVersion
text
contribution to journal
description <p dir="ltr">Researchers gravitate towards Generative Adversarial Networks (GAN) to create artificial images. However, GANs suffer from convergence issues, mode collapse, and overall complexity in balancing the Nash Equilibrium. Images generated are often distorted, rendering them useless. We propose a combination of Variational Autoencoders (VAEs) and a statistical oversampling method called K-Nearest Neighbor OveRsampling (KNNOR) to create artificial images. This combination of VAE and KNNOR results in more life-like images with reduced distortion. We fine-tune several pre-trained networks on a separate set of real and fake face images to test images generated by our method against images generated by conventional Deep Convolutional GANs (DCGANs). We also compare the combination of VAEs and Synthetic Minority Oversampling Technique (SMOTE) to establish the efficacy of KNNOR against naive oversampling methods. Not only are our methods better able to convince the classifiers that the images generated are authentic, but the models are also half in size of DCGANs. The code is available at GitHub for public use.</p><h2>Other Information</h2><p dir="ltr">Published in: IEEE Access<br>License: <a href="http://creativecommons.org/licenses/by/4.0" target="_blank">http://creativecommons.org/licenses/by/4.0</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/access.2023.3259236" target="_blank">https://dx.doi.org/10.1109/access.2023.3259236</a></p>
eu_rights_str_mv openAccess
id Manara2_a604c52ec5c08aa216b277485221cc3c
identifier_str_mv 10.1109/access.2023.3259236
network_acronym_str Manara2
network_name_str Manara2
oai_identifier_str oai:figshare.com:article/25204208
publishDate 2023
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling ApproachAshhadul Islam (16869981)Samir Brahim Belhaouari (9427347)EngineeringElectrical engineeringElectronics, sensors and digital hardwareMaterials engineeringGeneratorsFeature extractionGenerative adversarial networksDecodingTrainingNash equilibriumImage synthesisFace recognitionvariational autoencodersImage reconstructionartificial image creation<p dir="ltr">Researchers gravitate towards Generative Adversarial Networks (GAN) to create artificial images. However, GANs suffer from convergence issues, mode collapse, and overall complexity in balancing the Nash Equilibrium. Images generated are often distorted, rendering them useless. We propose a combination of Variational Autoencoders (VAEs) and a statistical oversampling method called K-Nearest Neighbor OveRsampling (KNNOR) to create artificial images. This combination of VAE and KNNOR results in more life-like images with reduced distortion. We fine-tune several pre-trained networks on a separate set of real and fake face images to test images generated by our method against images generated by conventional Deep Convolutional GANs (DCGANs). We also compare the combination of VAEs and Synthetic Minority Oversampling Technique (SMOTE) to establish the efficacy of KNNOR against naive oversampling methods. Not only are our methods better able to convince the classifiers that the images generated are authentic, but the models are also half in size of DCGANs. The code is available at GitHub for public use.</p><h2>Other Information</h2><p dir="ltr">Published in: IEEE Access<br>License: <a href="http://creativecommons.org/licenses/by/4.0" target="_blank">http://creativecommons.org/licenses/by/4.0</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/access.2023.3259236" target="_blank">https://dx.doi.org/10.1109/access.2023.3259236</a></p>2023-03-20T03:00:00ZTextJournal contributioninfo:eu-repo/semantics/publishedVersiontextcontribution to journal10.1109/access.2023.3259236https://figshare.com/articles/journal_contribution/Fast_and_Efficient_Image_Generation_Using_Variational_Autoencoders_and_K-Nearest_Neighbor_OveRsampling_Approach/25204208CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/252042082023-03-20T03:00:00Z
spellingShingle Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach
Ashhadul Islam (16869981)
Engineering
Electrical engineering
Electronics, sensors and digital hardware
Materials engineering
Generators
Feature extraction
Generative adversarial networks
Decoding
Training
Nash equilibrium
Image synthesis
Face recognition
variational autoencoders
Image reconstruction
artificial image creation
status_str publishedVersion
title Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach
title_full Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach
title_fullStr Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach
title_full_unstemmed Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach
title_short Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach
title_sort Fast and Efficient Image Generation Using Variational Autoencoders and K-Nearest Neighbor OveRsampling Approach
topic Engineering
Electrical engineering
Electronics, sensors and digital hardware
Materials engineering
Generators
Feature extraction
Generative adversarial networks
Decoding
Training
Nash equilibrium
Image synthesis
Face recognition
variational autoencoders
Image reconstruction
artificial image creation