How many specimens make a sufficient training set for automated 3D feature extraction?
Deep learning has emerged as a robust tool for automating feature extraction from 3D images, offering an efficient alternative to labour-intensive and potentially biased manual image segmentation methods. However, there has been limited exploration into the optimal training set sizes, including asse...
Main Authors: | , , , , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
2024
|
Subjects: | |
Online Access: | https://eprints.soton.ac.uk/490070/ https://eprints.soton.ac.uk/490070/1/J_Mulqueeney_RSOS-240113_Manuscript_Editable.docx |
id |
ftsouthampton:oai:eprints.soton.ac.uk:490070 |
---|---|
record_format |
openpolar |
spelling |
ftsouthampton:oai:eprints.soton.ac.uk:490070 2024-06-23T07:56:17+00:00 How many specimens make a sufficient training set for automated 3D feature extraction? Mulqueeney, James M. Searle-Barnes, Alex Brombacher, Anieke Sweeney, Marisa Goswami, Anjali Ezard, Thomas H.G. 2024-04-26 text https://eprints.soton.ac.uk/490070/ https://eprints.soton.ac.uk/490070/1/J_Mulqueeney_RSOS-240113_Manuscript_Editable.docx en English eng https://eprints.soton.ac.uk/490070/1/J_Mulqueeney_RSOS-240113_Manuscript_Editable.docx Mulqueeney, James M., Searle-Barnes, Alex, Brombacher, Anieke, Sweeney, Marisa, Goswami, Anjali and Ezard, Thomas H.G. (2024) How many specimens make a sufficient training set for automated 3D feature extraction? Royal Society Open Science. (In Press) cc_by_4 Article PeerReviewed 2024 ftsouthampton 2024-05-29T00:34:52Z Deep learning has emerged as a robust tool for automating feature extraction from 3D images, offering an efficient alternative to labour-intensive and potentially biased manual image segmentation methods. However, there has been limited exploration into the optimal training set sizes, including assessing whether artificial expansion by data augmentation can achieve consistent results in less time and how consistent these benefits are across different types of traits. In this study, we manually segmented 50 planktonic foraminifera specimens from the genus Menardella to determine the minimum number of training images required to produce accurate volumetric and shape data from internal and external structures. The results reveal unsurprisingly that deep learning models improve with a larger number of training images with eight specimens being required to achieve 95% accuracy. Furthermore, data augmentation can enhance network accuracy by up to 8.0%. Notably, predicting both volumetric and shape measurements for the internal structure poses a greater challenge compared to the external structure, due to low contrast between different materials and increased geometric complexity. These results provide novel insight into optimal training set sizes for precise image segmentation of diverse traits and highlight the potential of data augmentation for enhancing multivariate feature extraction from 3D images. Article in Journal/Newspaper Planktonic foraminifera University of Southampton: e-Prints Soton |
institution |
Open Polar |
collection |
University of Southampton: e-Prints Soton |
op_collection_id |
ftsouthampton |
language |
English |
description |
Deep learning has emerged as a robust tool for automating feature extraction from 3D images, offering an efficient alternative to labour-intensive and potentially biased manual image segmentation methods. However, there has been limited exploration into the optimal training set sizes, including assessing whether artificial expansion by data augmentation can achieve consistent results in less time and how consistent these benefits are across different types of traits. In this study, we manually segmented 50 planktonic foraminifera specimens from the genus Menardella to determine the minimum number of training images required to produce accurate volumetric and shape data from internal and external structures. The results reveal unsurprisingly that deep learning models improve with a larger number of training images with eight specimens being required to achieve 95% accuracy. Furthermore, data augmentation can enhance network accuracy by up to 8.0%. Notably, predicting both volumetric and shape measurements for the internal structure poses a greater challenge compared to the external structure, due to low contrast between different materials and increased geometric complexity. These results provide novel insight into optimal training set sizes for precise image segmentation of diverse traits and highlight the potential of data augmentation for enhancing multivariate feature extraction from 3D images. |
format |
Article in Journal/Newspaper |
author |
Mulqueeney, James M. Searle-Barnes, Alex Brombacher, Anieke Sweeney, Marisa Goswami, Anjali Ezard, Thomas H.G. |
spellingShingle |
Mulqueeney, James M. Searle-Barnes, Alex Brombacher, Anieke Sweeney, Marisa Goswami, Anjali Ezard, Thomas H.G. How many specimens make a sufficient training set for automated 3D feature extraction? |
author_facet |
Mulqueeney, James M. Searle-Barnes, Alex Brombacher, Anieke Sweeney, Marisa Goswami, Anjali Ezard, Thomas H.G. |
author_sort |
Mulqueeney, James M. |
title |
How many specimens make a sufficient training set for automated 3D feature extraction? |
title_short |
How many specimens make a sufficient training set for automated 3D feature extraction? |
title_full |
How many specimens make a sufficient training set for automated 3D feature extraction? |
title_fullStr |
How many specimens make a sufficient training set for automated 3D feature extraction? |
title_full_unstemmed |
How many specimens make a sufficient training set for automated 3D feature extraction? |
title_sort |
how many specimens make a sufficient training set for automated 3d feature extraction? |
publishDate |
2024 |
url |
https://eprints.soton.ac.uk/490070/ https://eprints.soton.ac.uk/490070/1/J_Mulqueeney_RSOS-240113_Manuscript_Editable.docx |
genre |
Planktonic foraminifera |
genre_facet |
Planktonic foraminifera |
op_relation |
https://eprints.soton.ac.uk/490070/1/J_Mulqueeney_RSOS-240113_Manuscript_Editable.docx Mulqueeney, James M., Searle-Barnes, Alex, Brombacher, Anieke, Sweeney, Marisa, Goswami, Anjali and Ezard, Thomas H.G. (2024) How many specimens make a sufficient training set for automated 3D feature extraction? Royal Society Open Science. (In Press) |
op_rights |
cc_by_4 |
_version_ |
1802649268941488128 |