基于 TC-UNet++ 和距离分水岭的细胞核图像分割和计数方法

IF 1.7 4区 医学 Q3 ENGINEERING, BIOMEDICAL
Kaifeng Zheng , Jie Pan , Ziyan Jia , Shuyan Xiao , Weige Tao , Dachuan Zhang , Qing Li , Lingjiao Pan
{"title":"基于 TC-UNet++ 和距离分水岭的细胞核图像分割和计数方法","authors":"Kaifeng Zheng ,&nbsp;Jie Pan ,&nbsp;Ziyan Jia ,&nbsp;Shuyan Xiao ,&nbsp;Weige Tao ,&nbsp;Dachuan Zhang ,&nbsp;Qing Li ,&nbsp;Lingjiao Pan","doi":"10.1016/j.medengphy.2024.104244","DOIUrl":null,"url":null,"abstract":"<div><div>Nucleus segmentation and counting play a crucial role in many cell analysis applications. However, the dense distribution and blurry boundaries of nucleus make nucleus segmentation tasks challenging. This paper proposes a novel segmentation and counting method. Firstly, TC-UNet++ is proposed to achieve a global segmentation. Then, the distance watershed method is used to finish local segmentation, which separate the adhesion and overlap part of the image. Finally, counting method is performed to obtain information on the counting number, area and center of mass of nucleus. TC-UNet++ achieved a Dice coefficient of 89.95% for cell instance segmentation on the Data Science Bowl dataset, surpassing the original U-Net++ by 0.23%. It also showed a 5.09% improvement in counting results compared to other methods. On the ALL-IDB dataset, TC-UNet++ reached a Dice coefficient of 83.97%, a 7.93% increase over the original U-Net++. Additionally, its counting results improved by 16.82% compared to other approaches. These results indicate that our method has a more complete and reasonable nucleus segmentation and counting scheme compared to other methods.</div></div>","PeriodicalId":49836,"journal":{"name":"Medical Engineering & Physics","volume":"133 ","pages":"Article 104244"},"PeriodicalIF":1.7000,"publicationDate":"2024-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A method of nucleus image segmentation and counting based on TC-UNet++ and distance watershed\",\"authors\":\"Kaifeng Zheng ,&nbsp;Jie Pan ,&nbsp;Ziyan Jia ,&nbsp;Shuyan Xiao ,&nbsp;Weige Tao ,&nbsp;Dachuan Zhang ,&nbsp;Qing Li ,&nbsp;Lingjiao Pan\",\"doi\":\"10.1016/j.medengphy.2024.104244\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Nucleus segmentation and counting play a crucial role in many cell analysis applications. However, the dense distribution and blurry boundaries of nucleus make nucleus segmentation tasks challenging. This paper proposes a novel segmentation and counting method. Firstly, TC-UNet++ is proposed to achieve a global segmentation. Then, the distance watershed method is used to finish local segmentation, which separate the adhesion and overlap part of the image. Finally, counting method is performed to obtain information on the counting number, area and center of mass of nucleus. TC-UNet++ achieved a Dice coefficient of 89.95% for cell instance segmentation on the Data Science Bowl dataset, surpassing the original U-Net++ by 0.23%. It also showed a 5.09% improvement in counting results compared to other methods. On the ALL-IDB dataset, TC-UNet++ reached a Dice coefficient of 83.97%, a 7.93% increase over the original U-Net++. Additionally, its counting results improved by 16.82% compared to other approaches. These results indicate that our method has a more complete and reasonable nucleus segmentation and counting scheme compared to other methods.</div></div>\",\"PeriodicalId\":49836,\"journal\":{\"name\":\"Medical Engineering & Physics\",\"volume\":\"133 \",\"pages\":\"Article 104244\"},\"PeriodicalIF\":1.7000,\"publicationDate\":\"2024-10-18\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Medical Engineering & Physics\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S1350453324001450\",\"RegionNum\":4,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"ENGINEERING, BIOMEDICAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Medical Engineering & Physics","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1350453324001450","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENGINEERING, BIOMEDICAL","Score":null,"Total":0}
引用次数: 0

摘要

细胞核的分割和计数在许多细胞分析应用中起着至关重要的作用。然而,由于细胞核分布密集且边界模糊,因此细胞核分割任务极具挑战性。本文提出了一种新颖的细胞核分割和计数方法。首先,提出了 TC-UNet++ 来实现全局分割。然后,使用距离分水岭方法完成局部分割,分离出图像的粘连和重叠部分。最后,采用计数法获得细胞核的计数数量、面积和质心等信息。在数据科学碗数据集上,TC-UNet++ 的细胞实例分割 Dice 系数达到 89.95%,比原始 U-Net++ 高出 0.23%。与其他方法相比,TC-Net++ 的计数结果提高了 5.09%。在 ALL-IDB 数据集上,TC-UNet++ 的骰子系数达到 83.97%,比原来的 U-Net++ 提高了 7.93%。此外,与其他方法相比,它的计数结果提高了 16.82%。这些结果表明,与其他方法相比,我们的方法具有更完整、更合理的核分割和计数方案。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
A method of nucleus image segmentation and counting based on TC-UNet++ and distance watershed
Nucleus segmentation and counting play a crucial role in many cell analysis applications. However, the dense distribution and blurry boundaries of nucleus make nucleus segmentation tasks challenging. This paper proposes a novel segmentation and counting method. Firstly, TC-UNet++ is proposed to achieve a global segmentation. Then, the distance watershed method is used to finish local segmentation, which separate the adhesion and overlap part of the image. Finally, counting method is performed to obtain information on the counting number, area and center of mass of nucleus. TC-UNet++ achieved a Dice coefficient of 89.95% for cell instance segmentation on the Data Science Bowl dataset, surpassing the original U-Net++ by 0.23%. It also showed a 5.09% improvement in counting results compared to other methods. On the ALL-IDB dataset, TC-UNet++ reached a Dice coefficient of 83.97%, a 7.93% increase over the original U-Net++. Additionally, its counting results improved by 16.82% compared to other approaches. These results indicate that our method has a more complete and reasonable nucleus segmentation and counting scheme compared to other methods.
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
Medical Engineering & Physics
Medical Engineering & Physics 工程技术-工程:生物医学
CiteScore
4.30
自引率
4.50%
发文量
172
审稿时长
3.0 months
期刊介绍: Medical Engineering & Physics provides a forum for the publication of the latest developments in biomedical engineering, and reflects the essential multidisciplinary nature of the subject. The journal publishes in-depth critical reviews, scientific papers and technical notes. Our focus encompasses the application of the basic principles of physics and engineering to the development of medical devices and technology, with the ultimate aim of producing improvements in the quality of health care.Topics covered include biomechanics, biomaterials, mechanobiology, rehabilitation engineering, biomedical signal processing and medical device development. Medical Engineering & Physics aims to keep both engineers and clinicians abreast of the latest applications of technology to health care.
×
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