Show simple item record

dc.contributor.author
Scebba, Gaetano
dc.contributor.author
Zhang, Jia
dc.contributor.author
Catanzaro, Sabrina
dc.contributor.author
Mihai, Carina
dc.contributor.author
Distler, Oliver
dc.contributor.author
Berli, Martin
dc.contributor.author
Karlen, Walter
dc.date.accessioned
2022-12-02T09:02:08Z
dc.date.available
2022-03-09T06:47:42Z
dc.date.available
2022-05-30T10:05:27Z
dc.date.available
2022-12-02T08:58:27Z
dc.date.available
2022-12-02T09:02:08Z
dc.date.issued
2022
dc.identifier.other
10.1016/j.imu.2022.100884
en_US
dc.identifier.uri
http://hdl.handle.net/20.500.11850/536186
dc.identifier.doi
10.3929/ethz-b-000536186
dc.description.abstract
Chronic wounds significantly impact quality of life. They can rapidly deteriorate and require close monitoring of healing progress. Image-based wound analysis is a way of objectively assessing the wound status by quantifying important features that are related to healing. However, high heterogeneity of the wound types and imaging conditions challenge the robust segmentation of wound images. We present Detect-and-Segment (DS), a deep learning approach to produce wound segmentation maps with high generalization capabilities. In our approach, dedicated deep neural networks detected the wound position, isolated the wound from the perturbing background, and computed a wound segmentation map. We tested this approach on a diabetic foot ulcers data set and compared it to a segmentation method based on the full image. To evaluate its generalizability on out-of-distribution data, we measured the performance of the DS approach on 4 additional independent data sets, with larger variety of wound types from different body locations. The Matthews’ correlation coefficient (MCC) improved from 0.29 (full image) to 0.85 (DS) on the diabetic foot ulcer data set. When the DS was tested on the independent data sets, the mean MCC increased from 0.17 to 0.85. Furthermore, the DS enabled the training of segmentation models with up to 90% less training data without impacting the segmentation performance. The proposed DS approach is a step towards automating wound analysis and reducing efforts to manage chronic wounds.
en_US
dc.format
application/pdf
en_US
dc.language.iso
en
en_US
dc.publisher
Elsevier
en_US
dc.rights.uri
http://creativecommons.org/licenses/by-nc-nd/4.0/
dc.subject
Chronic wounds
en_US
dc.subject
Semantic segmentation
en_US
dc.subject
Machine learning
en_US
dc.subject
Generalizability
en_US
dc.subject
Smartphone
en_US
dc.title
Detect-and-segment: A deep learning approach to automate wound image segmentation
en_US
dc.type
Journal Article
dc.rights.license
Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International
dc.date.published
2022-02-24
ethz.journal.title
Informatics in Medicine Unlocked
ethz.journal.volume
29
en_US
ethz.pages.start
100884
en_US
ethz.size
9 p.
en_US
ethz.version.deposit
publishedVersion
en_US
ethz.identifier.scopus
ethz.publication.place
Amsterdam
en_US
ethz.publication.status
published
en_US
ethz.date.deposited
2022-03-09T06:47:45Z
ethz.source
SCOPUS
ethz.eth
yes
en_US
ethz.availability
Open access
en_US
ethz.rosetta.installDate
2022-05-30T10:05:37Z
ethz.rosetta.lastUpdated
2023-02-07T08:11:49Z
ethz.rosetta.versionExported
true
ethz.COinS
ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.atitle=Detect-and-segment:%20A%20deep%20learning%20approach%20to%20automate%20wound%20image%20segmentation&rft.jtitle=Informatics%20in%20Medicine%20Unlocked&rft.date=2022&rft.volume=29&rft.spage=100884&rft.au=Scebba,%20Gaetano&Zhang,%20Jia&Catanzaro,%20Sabrina&Mihai,%20Carina&Distler,%20Oliver&rft.genre=article&rft_id=info:doi/10.1016/j.imu.2022.100884&
 Search print copy at ETH Library

Files in this item

Thumbnail

Publication type

Show simple item record