Cargando…

Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI

Deep learning (DL) models for segmenting various anatomical structures have achieved great success via a static DL model that is trained in a single source domain. Yet, the static DL model is likely to perform poorly in a continually evolving environment, requiring appropriate model updates. In an i...

Descripción completa

Detalles Bibliográficos
Autores principales: Liu, Xiaofeng, Shih, Helen A., Xing, Fangxu, Santarnecchi, Emiliano, El Fakhri, Georges, Woo, Jonghye
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Cornell University 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10312795/
https://www.ncbi.nlm.nih.gov/pubmed/37396599
_version_ 1785066987820417024
author Liu, Xiaofeng
Shih, Helen A.
Xing, Fangxu
Santarnecchi, Emiliano
El Fakhri, Georges
Woo, Jonghye
author_facet Liu, Xiaofeng
Shih, Helen A.
Xing, Fangxu
Santarnecchi, Emiliano
El Fakhri, Georges
Woo, Jonghye
author_sort Liu, Xiaofeng
collection PubMed
description Deep learning (DL) models for segmenting various anatomical structures have achieved great success via a static DL model that is trained in a single source domain. Yet, the static DL model is likely to perform poorly in a continually evolving environment, requiring appropriate model updates. In an incremental learning setting, we would expect that well-trained static models are updated, following continually evolving target domain data—e.g., additional lesions or structures of interest—collected from different sites, without catastrophic forgetting. This, however, poses challenges, due to distribution shifts, additional structures not seen during the initial model training, and the absence of training data in a source domain. To address these challenges, in this work, we seek to progressively evolve an “off-the-shelf” trained segmentation model to diverse datasets with additional anatomical categories in a unified manner. Specifically, we first propose a divergence-aware dual-flow module with balanced rigidity and plasticity branches to decouple old and new tasks, which is guided by continuous batch renormalization. Then, a complementary pseudo-label training scheme with self-entropy regularized momentum MixUp decay is developed for adaptive network optimization. We evaluated our framework on a brain tumor segmentation task with continually changing target domains—i.e., new MRI scanners/modalities with incremental structures. Our framework was able to well retain the discriminability of previously learned structures, hence enabling the realistic life-long segmentation model extension along with the widespread accumulation of big medical data.
format Online
Article
Text
id pubmed-10312795
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher Cornell University
record_format MEDLINE/PubMed
spelling pubmed-103127952023-07-01 Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI Liu, Xiaofeng Shih, Helen A. Xing, Fangxu Santarnecchi, Emiliano El Fakhri, Georges Woo, Jonghye ArXiv Article Deep learning (DL) models for segmenting various anatomical structures have achieved great success via a static DL model that is trained in a single source domain. Yet, the static DL model is likely to perform poorly in a continually evolving environment, requiring appropriate model updates. In an incremental learning setting, we would expect that well-trained static models are updated, following continually evolving target domain data—e.g., additional lesions or structures of interest—collected from different sites, without catastrophic forgetting. This, however, poses challenges, due to distribution shifts, additional structures not seen during the initial model training, and the absence of training data in a source domain. To address these challenges, in this work, we seek to progressively evolve an “off-the-shelf” trained segmentation model to diverse datasets with additional anatomical categories in a unified manner. Specifically, we first propose a divergence-aware dual-flow module with balanced rigidity and plasticity branches to decouple old and new tasks, which is guided by continuous batch renormalization. Then, a complementary pseudo-label training scheme with self-entropy regularized momentum MixUp decay is developed for adaptive network optimization. We evaluated our framework on a brain tumor segmentation task with continually changing target domains—i.e., new MRI scanners/modalities with incremental structures. Our framework was able to well retain the discriminability of previously learned structures, hence enabling the realistic life-long segmentation model extension along with the widespread accumulation of big medical data. Cornell University 2023-05-30 /pmc/articles/PMC10312795/ /pubmed/37396599 Text en https://creativecommons.org/licenses/by/4.0/This work is licensed under a Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/) , which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use.
spellingShingle Article
Liu, Xiaofeng
Shih, Helen A.
Xing, Fangxu
Santarnecchi, Emiliano
El Fakhri, Georges
Woo, Jonghye
Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI
title Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI
title_full Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI
title_fullStr Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI
title_full_unstemmed Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI
title_short Incremental Learning for Heterogeneous Structure Segmentation in Brain Tumor MRI
title_sort incremental learning for heterogeneous structure segmentation in brain tumor mri
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10312795/
https://www.ncbi.nlm.nih.gov/pubmed/37396599
work_keys_str_mv AT liuxiaofeng incrementallearningforheterogeneousstructuresegmentationinbraintumormri
AT shihhelena incrementallearningforheterogeneousstructuresegmentationinbraintumormri
AT xingfangxu incrementallearningforheterogeneousstructuresegmentationinbraintumormri
AT santarnecchiemiliano incrementallearningforheterogeneousstructuresegmentationinbraintumormri
AT elfakhrigeorges incrementallearningforheterogeneousstructuresegmentationinbraintumormri
AT woojonghye incrementallearningforheterogeneousstructuresegmentationinbraintumormri