Cargando…

Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images

In the application of deep learning on optical coherence tomography (OCT) data, it is common to train classification networks using 2D images originating from volumetric data. Given the micrometer resolution of OCT systems, consecutive images are often very similar in both visible structures and noi...

Descripción completa

Detalles Bibliográficos
Autores principales: Tampu, Iulian Emil, Eklund, Anders, Haj-Hosseini, Neda
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Nature Publishing Group UK 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9500039/
https://www.ncbi.nlm.nih.gov/pubmed/36138025
http://dx.doi.org/10.1038/s41597-022-01618-6
_version_ 1784795127453057024
author Tampu, Iulian Emil
Eklund, Anders
Haj-Hosseini, Neda
author_facet Tampu, Iulian Emil
Eklund, Anders
Haj-Hosseini, Neda
author_sort Tampu, Iulian Emil
collection PubMed
description In the application of deep learning on optical coherence tomography (OCT) data, it is common to train classification networks using 2D images originating from volumetric data. Given the micrometer resolution of OCT systems, consecutive images are often very similar in both visible structures and noise. Thus, an inappropriate data split can result in overlap between the training and testing sets, with a large portion of the literature overlooking this aspect. In this study, the effect of improper dataset splitting on model evaluation is demonstrated for three classification tasks using three OCT open-access datasets extensively used, Kermany’s and Srinivasan’s ophthalmology datasets, and AIIMS breast tissue dataset. Results show that the classification performance is inflated by 0.07 up to 0.43 in terms of Matthews Correlation Coefficient (accuracy: 5% to 30%) for models tested on datasets with improper splitting, highlighting the considerable effect of dataset handling on model evaluation. This study intends to raise awareness on the importance of dataset splitting given the increased research interest in implementing deep learning on OCT data.
format Online
Article
Text
id pubmed-9500039
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher Nature Publishing Group UK
record_format MEDLINE/PubMed
spelling pubmed-95000392022-09-24 Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images Tampu, Iulian Emil Eklund, Anders Haj-Hosseini, Neda Sci Data Analysis In the application of deep learning on optical coherence tomography (OCT) data, it is common to train classification networks using 2D images originating from volumetric data. Given the micrometer resolution of OCT systems, consecutive images are often very similar in both visible structures and noise. Thus, an inappropriate data split can result in overlap between the training and testing sets, with a large portion of the literature overlooking this aspect. In this study, the effect of improper dataset splitting on model evaluation is demonstrated for three classification tasks using three OCT open-access datasets extensively used, Kermany’s and Srinivasan’s ophthalmology datasets, and AIIMS breast tissue dataset. Results show that the classification performance is inflated by 0.07 up to 0.43 in terms of Matthews Correlation Coefficient (accuracy: 5% to 30%) for models tested on datasets with improper splitting, highlighting the considerable effect of dataset handling on model evaluation. This study intends to raise awareness on the importance of dataset splitting given the increased research interest in implementing deep learning on OCT data. Nature Publishing Group UK 2022-09-22 /pmc/articles/PMC9500039/ /pubmed/36138025 http://dx.doi.org/10.1038/s41597-022-01618-6 Text en © The Author(s) 2022 https://creativecommons.org/licenses/by/4.0/Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ (https://creativecommons.org/licenses/by/4.0/) .
spellingShingle Analysis
Tampu, Iulian Emil
Eklund, Anders
Haj-Hosseini, Neda
Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images
title Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images
title_full Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images
title_fullStr Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images
title_full_unstemmed Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images
title_short Inflation of test accuracy due to data leakage in deep learning-based classification of OCT images
title_sort inflation of test accuracy due to data leakage in deep learning-based classification of oct images
topic Analysis
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9500039/
https://www.ncbi.nlm.nih.gov/pubmed/36138025
http://dx.doi.org/10.1038/s41597-022-01618-6
work_keys_str_mv AT tampuiulianemil inflationoftestaccuracyduetodataleakageindeeplearningbasedclassificationofoctimages
AT eklundanders inflationoftestaccuracyduetodataleakageindeeplearningbasedclassificationofoctimages
AT hajhosseinineda inflationoftestaccuracyduetodataleakageindeeplearningbasedclassificationofoctimages