Multi-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning

Wound area segmentation really progressed with the emergence of deep learning, due to its robustness in uncontrolled lighting and no need to design hand-crafted features but two limits have still to be overcome: firstly, its performance relies on the size and quality of the training dataset in the m...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	R. Niri, E. Gutierrez, H. Douzi, Y. Lucas, S. Treuillet, B. Castaneda, I. Hernandez
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Deep learning 3D registration chronic wounds semantic segmentation data augmentation U-Net Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/734c0af4b81546ea83d4c1125658ac77
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:734c0af4b81546ea83d4c1125658ac77
record_format	dspace
spelling	oai:doaj.org-article:734c0af4b81546ea83d4c1125658ac772021-12-03T00:01:24ZMulti-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning2169-353610.1109/ACCESS.2021.3130784https://doaj.org/article/734c0af4b81546ea83d4c1125658ac772021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9627142/https://doaj.org/toc/2169-3536Wound area segmentation really progressed with the emergence of deep learning, due to its robustness in uncontrolled lighting and no need to design hand-crafted features but two limits have still to be overcome: firstly, its performance relies on the size and quality of the training dataset in the medical field, where data annotation is costly and time-consuming; secondly the accuracy of the segmentation depends highly on the camera distance and angle and moreover perspective effects prevent measuring real surfaces in single views. To address concurrently these two issues, we propose to apply multi-view modeling: an image sequence is acquired around the wound site and enables wound 3D reconstruction. Then, a segmentation step is run to extract roughly the wound from the background in each view and to select the best view with an original strategy. This view provides the most accurate segmentation and the real wound bed area even on non planar wounds. Finally, this segmentation is backprojected in each view to generate a complete set of well annotated real images to reinforce the learning step of the neural network. In our experiments, we compare several strategies to select the best view in the image sequence. The proposed method, tested on a dataset of 270 images, outperforms standard deep learning approach based on a single view, as recorded with DICE index and IoU score which rise respectively from 36.53% to 86.3% and 29.48% to 77.09% for the wound class to achieve an overall DICE and IoU score of 93.04% and 86.61% including background class. These results attest to the robustness of our method and its improved accuracy in the wound segmentation task.R. NiriE. GutierrezH. DouziY. LucasS. TreuilletB. CastanedaI. HernandezIEEEarticleDeep learning3D registrationchronic woundssemantic segmentationdata augmentationU-NetElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 157628-157638 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Deep learning 3D registration chronic wounds semantic segmentation data augmentation U-Net Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Deep learning 3D registration chronic wounds semantic segmentation data augmentation U-Net Electrical engineering. Electronics. Nuclear engineering TK1-9971 R. Niri E. Gutierrez H. Douzi Y. Lucas S. Treuillet B. Castaneda I. Hernandez Multi-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning
description	Wound area segmentation really progressed with the emergence of deep learning, due to its robustness in uncontrolled lighting and no need to design hand-crafted features but two limits have still to be overcome: firstly, its performance relies on the size and quality of the training dataset in the medical field, where data annotation is costly and time-consuming; secondly the accuracy of the segmentation depends highly on the camera distance and angle and moreover perspective effects prevent measuring real surfaces in single views. To address concurrently these two issues, we propose to apply multi-view modeling: an image sequence is acquired around the wound site and enables wound 3D reconstruction. Then, a segmentation step is run to extract roughly the wound from the background in each view and to select the best view with an original strategy. This view provides the most accurate segmentation and the real wound bed area even on non planar wounds. Finally, this segmentation is backprojected in each view to generate a complete set of well annotated real images to reinforce the learning step of the neural network. In our experiments, we compare several strategies to select the best view in the image sequence. The proposed method, tested on a dataset of 270 images, outperforms standard deep learning approach based on a single view, as recorded with DICE index and IoU score which rise respectively from 36.53% to 86.3% and 29.48% to 77.09% for the wound class to achieve an overall DICE and IoU score of 93.04% and 86.61% including background class. These results attest to the robustness of our method and its improved accuracy in the wound segmentation task.
format	article
author	R. Niri E. Gutierrez H. Douzi Y. Lucas S. Treuillet B. Castaneda I. Hernandez
author_facet	R. Niri E. Gutierrez H. Douzi Y. Lucas S. Treuillet B. Castaneda I. Hernandez
author_sort	R. Niri
title	Multi-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning
title_short	Multi-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning
title_full	Multi-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning
title_fullStr	Multi-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning
title_full_unstemmed	Multi-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning
title_sort	multi-view data augmentation to improve wound segmentation on 3d surface model by deep learning
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/734c0af4b81546ea83d4c1125658ac77
work_keys_str_mv	AT rniri multiviewdataaugmentationtoimprovewoundsegmentationon3dsurfacemodelbydeeplearning AT egutierrez multiviewdataaugmentationtoimprovewoundsegmentationon3dsurfacemodelbydeeplearning AT hdouzi multiviewdataaugmentationtoimprovewoundsegmentationon3dsurfacemodelbydeeplearning AT ylucas multiviewdataaugmentationtoimprovewoundsegmentationon3dsurfacemodelbydeeplearning AT streuillet multiviewdataaugmentationtoimprovewoundsegmentationon3dsurfacemodelbydeeplearning AT bcastaneda multiviewdataaugmentationtoimprovewoundsegmentationon3dsurfacemodelbydeeplearning AT ihernandez multiviewdataaugmentationtoimprovewoundsegmentationon3dsurfacemodelbydeeplearning
_version_	1718373994243031040

Multi-View Data Augmentation to Improve Wound Segmentation on 3D Surface Model by Deep Learning

Ejemplares similares