Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Preliminary evaluation of an automated autoencoder-UNet pipeline for chemical image segmentation and compression with reference to serial ground truth pathology

Rafsanjani, Mohd Rifqi ; McBrien, Thomas ; Jirstrom, Karin LU orcid ; Rahman, Arman ; Prehn, Jochen H.M. ; Gallagher, William and Meade, Aidan D. (2024) Data Science for Photonics and Biophotonics 2024 In Proceedings of SPIE - The International Society for Optical Engineering 13011.
Abstract

The rapid advancement of imaging technologies in pathology has ushered in an era of data-intensive diagnostic workflows, generating large volumes of data that demand sophisticated segmentation and compression techniques. Chemical imaging approaches offer an all-digital objective approach to pathological analysis, though image segmentation is required for efficient computation. Convolutional autoencoders are highly connected deep learning networks which can learn salient features within imaging data for the purposes of compression, data recovery, development of classifiers and/or segmentation. In this study an objective analysis of a U-Net convolutional autoencoders for unsupervised image segmentation is conducted with respect to... (More)

The rapid advancement of imaging technologies in pathology has ushered in an era of data-intensive diagnostic workflows, generating large volumes of data that demand sophisticated segmentation and compression techniques. Chemical imaging approaches offer an all-digital objective approach to pathological analysis, though image segmentation is required for efficient computation. Convolutional autoencoders are highly connected deep learning networks which can learn salient features within imaging data for the purposes of compression, data recovery, development of classifiers and/or segmentation. In this study an objective analysis of a U-Net convolutional autoencoders for unsupervised image segmentation is conducted with respect to haematoxylin-eosin based ground-truth diagnostic pathology. We find that a light-weight network architecture may provide a suitable segmentation approach for chemical imaging.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
autoencoder-UNet, Breast cancer, Fourier Transform Infrared (FTIR) chemical imaging, image segmentation
host publication
Data Science for Photonics and Biophotonics
series title
Proceedings of SPIE - The International Society for Optical Engineering
editor
Bocklitz, Thomas
volume
13011
article number
130110D
publisher
SPIE
conference name
Data Science for Photonics and Biophotonics 2024
conference location
Strasbourg, France
conference dates
2024-04-07 - 2024-04-11
external identifiers
  • scopus:85200262180
ISSN
0277-786X
1996-756X
ISBN
9781510673403
DOI
10.1117/12.3022279
language
English
LU publication?
yes
id
bf426952-39d9-445c-854b-857bfb2e8313
date added to LUP
2024-11-11 14:17:37
date last changed
2025-05-27 06:10:15
@inproceedings{bf426952-39d9-445c-854b-857bfb2e8313,
  abstract     = {{<p>The rapid advancement of imaging technologies in pathology has ushered in an era of data-intensive diagnostic workflows, generating large volumes of data that demand sophisticated segmentation and compression techniques. Chemical imaging approaches offer an all-digital objective approach to pathological analysis, though image segmentation is required for efficient computation. Convolutional autoencoders are highly connected deep learning networks which can learn salient features within imaging data for the purposes of compression, data recovery, development of classifiers and/or segmentation. In this study an objective analysis of a U-Net convolutional autoencoders for unsupervised image segmentation is conducted with respect to haematoxylin-eosin based ground-truth diagnostic pathology. We find that a light-weight network architecture may provide a suitable segmentation approach for chemical imaging.</p>}},
  author       = {{Rafsanjani, Mohd Rifqi and McBrien, Thomas and Jirstrom, Karin and Rahman, Arman and Prehn, Jochen H.M. and Gallagher, William and Meade, Aidan D.}},
  booktitle    = {{Data Science for Photonics and Biophotonics}},
  editor       = {{Bocklitz, Thomas}},
  isbn         = {{9781510673403}},
  issn         = {{0277-786X}},
  keywords     = {{autoencoder-UNet; Breast cancer; Fourier Transform Infrared (FTIR) chemical imaging; image segmentation}},
  language     = {{eng}},
  publisher    = {{SPIE}},
  series       = {{Proceedings of SPIE - The International Society for Optical Engineering}},
  title        = {{Preliminary evaluation of an automated autoencoder-UNet pipeline for chemical image segmentation and compression with reference to serial ground truth pathology}},
  url          = {{http://dx.doi.org/10.1117/12.3022279}},
  doi          = {{10.1117/12.3022279}},
  volume       = {{13011}},
  year         = {{2024}},
}