Deep-learning-based image compression for microscopy images: An empirical study

With the fast development of modern microscopes and bioimaging techniques, an unprecedentedly large amount of imaging data is being generated, stored, analyzed, and shared through networks. The size of the data poses great challenges for current data infrastructure. One common way to reduce the data...

Full description

Saved in:
Bibliographic Details
Main Authors: Yu Zhou, Jan Sollmann, Jianxu Chen
Format: Article
Language:English
Published: Cambridge University Press 2024-01-01
Series:Biological Imaging
Subjects:
Online Access:https://www.cambridge.org/core/product/identifier/S2633903X24000151/type/journal_article
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846114658564440064
author Yu Zhou
Jan Sollmann
Jianxu Chen
author_facet Yu Zhou
Jan Sollmann
Jianxu Chen
author_sort Yu Zhou
collection DOAJ
description With the fast development of modern microscopes and bioimaging techniques, an unprecedentedly large amount of imaging data is being generated, stored, analyzed, and shared through networks. The size of the data poses great challenges for current data infrastructure. One common way to reduce the data size is by image compression. This study analyzes multiple classic and deep-learning-based image compression methods, as well as an empirical study on their impact on downstream deep-learning-based image processing models. We used deep-learning-based label-free prediction models (i.e., predicting fluorescent images from bright-field images) as an example downstream task for the comparison and analysis of the impact of image compression. Different compression techniques are compared in compression ratio, image similarity, and, most importantly, the prediction accuracy of label-free models on original and compressed images. We found that artificial intelligence (AI)-based compression techniques largely outperform the classic ones with minimal influence on the downstream 2D label-free tasks. In the end, we hope this study could shed light on the potential of deep-learning-based image compression and raise the awareness of the potential impacts of image compression on downstream deep-learning models for analysis.
format Article
id doaj-art-476c1627d2bb40e3abc4622734b2fdb8
institution Kabale University
issn 2633-903X
language English
publishDate 2024-01-01
publisher Cambridge University Press
record_format Article
series Biological Imaging
spelling doaj-art-476c1627d2bb40e3abc4622734b2fdb82024-12-20T09:02:59ZengCambridge University PressBiological Imaging2633-903X2024-01-01410.1017/S2633903X24000151Deep-learning-based image compression for microscopy images: An empirical studyYu Zhou0https://orcid.org/0009-0002-3914-1102Jan Sollmann1https://orcid.org/0009-0007-3973-4641Jianxu Chen2https://orcid.org/0000-0002-8500-1357Department of Biospectroscopy, Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V., Dortmund, Germany Faculty of Computer Science, Ruhr University Bochum, Bochum, GermanyDepartment of Biospectroscopy, Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V., Dortmund, Germany Faculty of Computer Science, Ruhr University Bochum, Bochum, GermanyDepartment of Biospectroscopy, Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V., Dortmund, GermanyWith the fast development of modern microscopes and bioimaging techniques, an unprecedentedly large amount of imaging data is being generated, stored, analyzed, and shared through networks. The size of the data poses great challenges for current data infrastructure. One common way to reduce the data size is by image compression. This study analyzes multiple classic and deep-learning-based image compression methods, as well as an empirical study on their impact on downstream deep-learning-based image processing models. We used deep-learning-based label-free prediction models (i.e., predicting fluorescent images from bright-field images) as an example downstream task for the comparison and analysis of the impact of image compression. Different compression techniques are compared in compression ratio, image similarity, and, most importantly, the prediction accuracy of label-free models on original and compressed images. We found that artificial intelligence (AI)-based compression techniques largely outperform the classic ones with minimal influence on the downstream 2D label-free tasks. In the end, we hope this study could shed light on the potential of deep-learning-based image compression and raise the awareness of the potential impacts of image compression on downstream deep-learning models for analysis.https://www.cambridge.org/core/product/identifier/S2633903X24000151/type/journal_articlecompressiondeep learningin-silico labellingmicroscopic images
spellingShingle Yu Zhou
Jan Sollmann
Jianxu Chen
Deep-learning-based image compression for microscopy images: An empirical study
Biological Imaging
compression
deep learning
in-silico labelling
microscopic images
title Deep-learning-based image compression for microscopy images: An empirical study
title_full Deep-learning-based image compression for microscopy images: An empirical study
title_fullStr Deep-learning-based image compression for microscopy images: An empirical study
title_full_unstemmed Deep-learning-based image compression for microscopy images: An empirical study
title_short Deep-learning-based image compression for microscopy images: An empirical study
title_sort deep learning based image compression for microscopy images an empirical study
topic compression
deep learning
in-silico labelling
microscopic images
url https://www.cambridge.org/core/product/identifier/S2633903X24000151/type/journal_article
work_keys_str_mv AT yuzhou deeplearningbasedimagecompressionformicroscopyimagesanempiricalstudy
AT jansollmann deeplearningbasedimagecompressionformicroscopyimagesanempiricalstudy
AT jianxuchen deeplearningbasedimagecompressionformicroscopyimagesanempiricalstudy