Target-aware cross-modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentation

Abstract There is growing interest in research on segmentation for the vestibular schwannoma (VS) and cochlea using high-resolution T2 (hrT2) imaging over contrast-enhanced T1 (ceT1) imaging due to the contrast agent side effects. However, the hrT2 imaging remains a problem of insufficient annotated...

Full description

Saved in:
Bibliographic Details
Main Authors: Bogyeong Kang, Hyeonyeong Nam, Myeongkyun Kang, Keun-Soo Heo, Minjoo Lim, Ji-Hye Oh, Tae-Eui Kam
Format: Article
Language:English
Published: Nature Portfolio 2024-11-01
Series:Scientific Reports
Online Access:https://doi.org/10.1038/s41598-024-77633-x
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846165316097277952
author Bogyeong Kang
Hyeonyeong Nam
Myeongkyun Kang
Keun-Soo Heo
Minjoo Lim
Ji-Hye Oh
Tae-Eui Kam
author_facet Bogyeong Kang
Hyeonyeong Nam
Myeongkyun Kang
Keun-Soo Heo
Minjoo Lim
Ji-Hye Oh
Tae-Eui Kam
author_sort Bogyeong Kang
collection DOAJ
description Abstract There is growing interest in research on segmentation for the vestibular schwannoma (VS) and cochlea using high-resolution T2 (hrT2) imaging over contrast-enhanced T1 (ceT1) imaging due to the contrast agent side effects. However, the hrT2 imaging remains a problem of insufficient annotated data, which is fatal for building more robust segmentation models. To address the issue, recent studies have adopted unsupervised domain adaptation approaches that translate ceT1 images to hrT2 images. However, previous studies did not consider the size and visual characteristics of the target objects, such as VS and cochlea, during image translation. Specifically, those works simply performed normalization on the entire image without considering its significant impact on the quality of the translated images. These approaches tend to erase the small target objects, making it difficult to preserve the structure of these objects when generating pseudo-target images. Furthermore, they may also struggle to accurately reflect the unique style of the target objects within the images. Therefore, we propose a target-aware unsupervised domain adaptation framework, designed for translating target objects, each tailored to their unique visual characteristics and size using target-aware normalization. We demonstrate the superiority of the proposed framework on a publicly available challenge dataset. Codes are available at https://github.com/Bokyeong-Kang/TANQ .
format Article
id doaj-art-37106a6db6f84a5eb0fea6db9ceef1e6
institution Kabale University
issn 2045-2322
language English
publishDate 2024-11-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-37106a6db6f84a5eb0fea6db9ceef1e62024-11-17T12:21:15ZengNature PortfolioScientific Reports2045-23222024-11-0114111210.1038/s41598-024-77633-xTarget-aware cross-modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentationBogyeong Kang0Hyeonyeong Nam1Myeongkyun Kang2Keun-Soo Heo3Minjoo Lim4Ji-Hye Oh5Tae-Eui Kam6Department of Artificial Intelligence, Korea UniversityDepartment of Artificial Intelligence, Korea UniversityDepartment of Robotics and Mechatronics Engineering, Daegu Gyeongbuk Institute of Science and Technology (DGIST)Department of Artificial Intelligence, Korea UniversityDepartment of Artificial Intelligence, Korea UniversityDepartment of Artificial Intelligence, Korea UniversityDepartment of Artificial Intelligence, Korea UniversityAbstract There is growing interest in research on segmentation for the vestibular schwannoma (VS) and cochlea using high-resolution T2 (hrT2) imaging over contrast-enhanced T1 (ceT1) imaging due to the contrast agent side effects. However, the hrT2 imaging remains a problem of insufficient annotated data, which is fatal for building more robust segmentation models. To address the issue, recent studies have adopted unsupervised domain adaptation approaches that translate ceT1 images to hrT2 images. However, previous studies did not consider the size and visual characteristics of the target objects, such as VS and cochlea, during image translation. Specifically, those works simply performed normalization on the entire image without considering its significant impact on the quality of the translated images. These approaches tend to erase the small target objects, making it difficult to preserve the structure of these objects when generating pseudo-target images. Furthermore, they may also struggle to accurately reflect the unique style of the target objects within the images. Therefore, we propose a target-aware unsupervised domain adaptation framework, designed for translating target objects, each tailored to their unique visual characteristics and size using target-aware normalization. We demonstrate the superiority of the proposed framework on a publicly available challenge dataset. Codes are available at https://github.com/Bokyeong-Kang/TANQ .https://doi.org/10.1038/s41598-024-77633-x
spellingShingle Bogyeong Kang
Hyeonyeong Nam
Myeongkyun Kang
Keun-Soo Heo
Minjoo Lim
Ji-Hye Oh
Tae-Eui Kam
Target-aware cross-modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentation
Scientific Reports
title Target-aware cross-modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentation
title_full Target-aware cross-modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentation
title_fullStr Target-aware cross-modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentation
title_full_unstemmed Target-aware cross-modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentation
title_short Target-aware cross-modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentation
title_sort target aware cross modality unsupervised domain adaptation for vestibular schwannoma and cochlea segmentation
url https://doi.org/10.1038/s41598-024-77633-x
work_keys_str_mv AT bogyeongkang targetawarecrossmodalityunsuperviseddomainadaptationforvestibularschwannomaandcochleasegmentation
AT hyeonyeongnam targetawarecrossmodalityunsuperviseddomainadaptationforvestibularschwannomaandcochleasegmentation
AT myeongkyunkang targetawarecrossmodalityunsuperviseddomainadaptationforvestibularschwannomaandcochleasegmentation
AT keunsooheo targetawarecrossmodalityunsuperviseddomainadaptationforvestibularschwannomaandcochleasegmentation
AT minjoolim targetawarecrossmodalityunsuperviseddomainadaptationforvestibularschwannomaandcochleasegmentation
AT jihyeoh targetawarecrossmodalityunsuperviseddomainadaptationforvestibularschwannomaandcochleasegmentation
AT taeeuikam targetawarecrossmodalityunsuperviseddomainadaptationforvestibularschwannomaandcochleasegmentation