扩张多尺度残留注意 U-Net:用于脑肿瘤分割的三维(3D)扩张多尺度残留注意 U-Net。

IF 2.9 2区 医学 Q2 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING
Quantitative Imaging in Medicine and Surgery Pub Date : 2024-10-01 Epub Date: 2024-09-19 DOI:10.21037/qims-24-779
Lihong Zhang, Yuzhuo Li, Yingbo Liang, Chongxin Xu, Tong Liu, Junding Sun
{"title":"扩张多尺度残留注意 U-Net:用于脑肿瘤分割的三维(3D)扩张多尺度残留注意 U-Net。","authors":"Lihong Zhang, Yuzhuo Li, Yingbo Liang, Chongxin Xu, Tong Liu, Junding Sun","doi":"10.21037/qims-24-779","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>The precise identification of the position and form of a tumor mass can improve early diagnosis and treatment. However, due to the complicated tumor categories and varying sizes and forms, the segregation of brain gliomas and their internal sub-regions is still very challenging. This study sought to design a new deep-learning network based on three-dimensional (3D) U-Net to address its shortcomings in brain tumor segmentation (BraTS) tasks.</p><p><strong>Methods: </strong>We developed a 3D dilated multi-scale residual attention U-Net (DMRA-U-Net) model for magnetic resonance imaging (MRI) BraTS. It used dilated convolution residual (DCR) modules to better process shallow features, multi-scale convolution residual (MCR) modules in the bottom encoding path to create richer and more comprehensive feature expression while reducing overall information loss or blurring, and a channel attention (CA) module between the encoding and decoding paths to address the problem of retrieving and preserving important features during the processing of deep feature maps.</p><p><strong>Results: </strong>The BraTS 2018-2021 datasets served as the training and evaluation datasets for this study. Further, the proposed architecture was assessed using metrics such as the dice similarity coefficient (DSC), Hausdorff distance (HD), and sensitivity (Sens). The DMRA U-Net model segments the whole tumor (WT), and the tumor core (TC), and the enhancing tumor (ET) regions of brain tumors. Using the suggested architecture, the DSCs were 0.9012, 0.8867, and 0.8813, the HDs were 28.86, 13.34, and 10.88 mm, and the Sens was 0.9429, 0.9452, and 0.9303 for the WT, TC, and ET regions, respectively. Compared to the traditional 3D U-Net, the DSC of the DMRA U-Net increased by 4.5%, 2.5%, and 0.8%, the HD of the DMRA U-Net decreased by 21.83, 16.42, and 10.00, the Sens of the DMRA U-Net increased by 0.4%, 0.7%, and 1.4% for the WT, TC, and ET regions, respectively. Further, the results of the statistical comparison of the performance indicators revealed that our model performed well generally in the segmentation of the WT, TC, and ET regions.</p><p><strong>Conclusions: </strong>We developed a promising tumor segmentation model. Our solution is open sourced and is available at: https://github.com/Gold3nk/dmra-unet.</p>","PeriodicalId":54267,"journal":{"name":"Quantitative Imaging in Medicine and Surgery","volume":null,"pages":null},"PeriodicalIF":2.9000,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11485367/pdf/","citationCount":"0","resultStr":"{\"title\":\"Dilated multi-scale residual attention (DMRA) U-Net: three-dimensional (3D) dilated multi-scale residual attention U-Net for brain tumor segmentation.\",\"authors\":\"Lihong Zhang, Yuzhuo Li, Yingbo Liang, Chongxin Xu, Tong Liu, Junding Sun\",\"doi\":\"10.21037/qims-24-779\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background: </strong>The precise identification of the position and form of a tumor mass can improve early diagnosis and treatment. However, due to the complicated tumor categories and varying sizes and forms, the segregation of brain gliomas and their internal sub-regions is still very challenging. This study sought to design a new deep-learning network based on three-dimensional (3D) U-Net to address its shortcomings in brain tumor segmentation (BraTS) tasks.</p><p><strong>Methods: </strong>We developed a 3D dilated multi-scale residual attention U-Net (DMRA-U-Net) model for magnetic resonance imaging (MRI) BraTS. It used dilated convolution residual (DCR) modules to better process shallow features, multi-scale convolution residual (MCR) modules in the bottom encoding path to create richer and more comprehensive feature expression while reducing overall information loss or blurring, and a channel attention (CA) module between the encoding and decoding paths to address the problem of retrieving and preserving important features during the processing of deep feature maps.</p><p><strong>Results: </strong>The BraTS 2018-2021 datasets served as the training and evaluation datasets for this study. Further, the proposed architecture was assessed using metrics such as the dice similarity coefficient (DSC), Hausdorff distance (HD), and sensitivity (Sens). The DMRA U-Net model segments the whole tumor (WT), and the tumor core (TC), and the enhancing tumor (ET) regions of brain tumors. Using the suggested architecture, the DSCs were 0.9012, 0.8867, and 0.8813, the HDs were 28.86, 13.34, and 10.88 mm, and the Sens was 0.9429, 0.9452, and 0.9303 for the WT, TC, and ET regions, respectively. Compared to the traditional 3D U-Net, the DSC of the DMRA U-Net increased by 4.5%, 2.5%, and 0.8%, the HD of the DMRA U-Net decreased by 21.83, 16.42, and 10.00, the Sens of the DMRA U-Net increased by 0.4%, 0.7%, and 1.4% for the WT, TC, and ET regions, respectively. Further, the results of the statistical comparison of the performance indicators revealed that our model performed well generally in the segmentation of the WT, TC, and ET regions.</p><p><strong>Conclusions: </strong>We developed a promising tumor segmentation model. Our solution is open sourced and is available at: https://github.com/Gold3nk/dmra-unet.</p>\",\"PeriodicalId\":54267,\"journal\":{\"name\":\"Quantitative Imaging in Medicine and Surgery\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":2.9000,\"publicationDate\":\"2024-10-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11485367/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Quantitative Imaging in Medicine and Surgery\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.21037/qims-24-779\",\"RegionNum\":2,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2024/9/19 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"Q2\",\"JCRName\":\"RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Quantitative Imaging in Medicine and Surgery","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.21037/qims-24-779","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/9/19 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING","Score":null,"Total":0}
引用次数: 0

摘要

背景:准确识别肿瘤肿块的位置和形态可以提高早期诊断和治疗的效果。然而,由于肿瘤类别复杂、大小形态各异,脑胶质瘤及其内部亚区域的分离仍然非常具有挑战性。本研究试图设计一种基于三维(3D)U-Net的新型深度学习网络,以解决其在脑肿瘤分割(BraTS)任务中的不足:我们为磁共振成像(MRI)BraTS开发了一个三维扩张多尺度残差注意U-Net(DMRA-U-Net)模型。它使用扩张卷积残差(DCR)模块来更好地处理浅层特征,在底部编码路径中使用多尺度卷积残差(MCR)模块来创建更丰富、更全面的特征表达,同时减少整体信息丢失或模糊,在编码和解码路径之间使用通道注意(CA)模块来解决在处理深层特征图时检索和保留重要特征的问题:BraTS 2018-2021 数据集是本研究的训练和评估数据集。此外,还使用骰子相似系数(DSC)、豪斯多夫距离(HD)和灵敏度(Sens)等指标对所提出的架构进行了评估。DMRA U-Net 模型可分割脑肿瘤的整个肿瘤(WT)、肿瘤核心(TC)和增强肿瘤(ET)区域。使用建议的结构,WT、TC 和 ET 区域的 DSC 分别为 0.9012、0.8867 和 0.8813,HD 分别为 28.86、13.34 和 10.88 mm,Sens 分别为 0.9429、0.9452 和 0.9303。与传统的三维 U-Net 相比,DMRA U-Net 的 DSC 分别增加了 4.5%、2.5% 和 0.8%,DMRA U-Net 的 HD 分别减少了 21.83、16.42 和 10.00,DMRA U-Net 的 Sens 在 WT、TC 和 ET 区域分别增加了 0.4%、0.7% 和 1.4%。此外,性能指标的统计比较结果表明,我们的模型在分割 WT、TC 和 ET 区域时总体表现良好:我们开发出了一种很有前景的肿瘤分割模型。我们的解决方案是开源的,可在以下网址获取:https://github.com/Gold3nk/dmra-unet.
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Dilated multi-scale residual attention (DMRA) U-Net: three-dimensional (3D) dilated multi-scale residual attention U-Net for brain tumor segmentation.

Background: The precise identification of the position and form of a tumor mass can improve early diagnosis and treatment. However, due to the complicated tumor categories and varying sizes and forms, the segregation of brain gliomas and their internal sub-regions is still very challenging. This study sought to design a new deep-learning network based on three-dimensional (3D) U-Net to address its shortcomings in brain tumor segmentation (BraTS) tasks.

Methods: We developed a 3D dilated multi-scale residual attention U-Net (DMRA-U-Net) model for magnetic resonance imaging (MRI) BraTS. It used dilated convolution residual (DCR) modules to better process shallow features, multi-scale convolution residual (MCR) modules in the bottom encoding path to create richer and more comprehensive feature expression while reducing overall information loss or blurring, and a channel attention (CA) module between the encoding and decoding paths to address the problem of retrieving and preserving important features during the processing of deep feature maps.

Results: The BraTS 2018-2021 datasets served as the training and evaluation datasets for this study. Further, the proposed architecture was assessed using metrics such as the dice similarity coefficient (DSC), Hausdorff distance (HD), and sensitivity (Sens). The DMRA U-Net model segments the whole tumor (WT), and the tumor core (TC), and the enhancing tumor (ET) regions of brain tumors. Using the suggested architecture, the DSCs were 0.9012, 0.8867, and 0.8813, the HDs were 28.86, 13.34, and 10.88 mm, and the Sens was 0.9429, 0.9452, and 0.9303 for the WT, TC, and ET regions, respectively. Compared to the traditional 3D U-Net, the DSC of the DMRA U-Net increased by 4.5%, 2.5%, and 0.8%, the HD of the DMRA U-Net decreased by 21.83, 16.42, and 10.00, the Sens of the DMRA U-Net increased by 0.4%, 0.7%, and 1.4% for the WT, TC, and ET regions, respectively. Further, the results of the statistical comparison of the performance indicators revealed that our model performed well generally in the segmentation of the WT, TC, and ET regions.

Conclusions: We developed a promising tumor segmentation model. Our solution is open sourced and is available at: https://github.com/Gold3nk/dmra-unet.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Quantitative Imaging in Medicine and Surgery
Quantitative Imaging in Medicine and Surgery Medicine-Radiology, Nuclear Medicine and Imaging
CiteScore
4.20
自引率
17.90%
发文量
252
期刊介绍: Information not localized
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信