Automatic segmentation of the prostate on MR images based on anatomy and deep learning

Lei Tao; Ling Ma; Maoqiang Xie; Xiabi Liu; Zhiqiang Tian; Baowei Fei

doi:10.1117/12.2581893

Automatic segmentation of the prostate on MR images based on anatomy and deep learning

Lei Tao
, Ling Ma^*
, Maoqiang Xie
, Xiabi Liu
, Zhiqiang Tian
, Baowei Fei

^*此作品的通讯作者

计算机学院

Nankai University
Xi'an Jiaotong University
University of Texas at Dallas
University of Texas Southwestern Medical Center

科研成果: 书/报告/会议事项章节 › 会议稿件 › 同行评审

7 引用（Scopus）

摘要

Accurate segmentation of the prostate has many applications in the detection, diagnosis and treatment of prostate cancer. Automatic segmentation can be a challenging task because of the inhomogeneous intensity distributions on MR images. In this paper, we propose an automatic segmentation method for the prostate on MR images based on anatomy. We use the 3D U-Net guided by anatomy knowledge, including the location and shape prior knowledge of the prostate on MR images, to constrain the segmentation of the gland. The proposed method has been evaluated on the public dataset PROMISE2012. Experimental results show that the proposed method achieves a mean Dice similarity coefficient of 91.6% as compared to the manual segmentation. The experimental results indicate that the proposed method based on anatomy knowledge can achieve satisfactory segmentation performance for prostate MRI.

源语言	英语
主期刊名	Medical Imaging 2021
主期刊副标题	Image-Guided Procedures, Robotic Interventions, and Modeling
编辑	Cristian A. Linte, Jeffrey H. Siewerdsen
出版商	SPIE
ISBN（电子版）	9781510640252
DOI	https://doi.org/10.1117/12.2581893
出版状态	已出版 - 2021
活动	Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling - Virtual, Online 期限: 15 2月 2021 → 19 2月 2021

出版系列

姓名	Progress in Biomedical Optics and Imaging - Proceedings of SPIE
卷	11598
ISSN（印刷版）	1605-7422

会议

会议	Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling
市	Virtual, Online
时期	15/02/21 → 19/02/21

联合国可持续发展目标

此成果有助于实现下列可持续发展目标：

可持续发展目标 3 良好健康与福祉

访问文件

10.1117/12.2581893

其它文件与链接

链接到 Scopus 的出版物

引用此

Tao, L., Ma, L., Xie, M., Liu, X., Tian, Z., & Fei, B. (2021). Automatic segmentation of the prostate on MR images based on anatomy and deep learning. 在 C. A. Linte, & J. H. Siewerdsen (编辑), Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling 文章 115981N (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; 卷 11598). SPIE. https://doi.org/10.1117/12.2581893

@inproceedings{fb7ef7f4ed7a4a288c62a243dda6a8ab,

title = "Automatic segmentation of the prostate on MR images based on anatomy and deep learning",

abstract = "Accurate segmentation of the prostate has many applications in the detection, diagnosis and treatment of prostate cancer. Automatic segmentation can be a challenging task because of the inhomogeneous intensity distributions on MR images. In this paper, we propose an automatic segmentation method for the prostate on MR images based on anatomy. We use the 3D U-Net guided by anatomy knowledge, including the location and shape prior knowledge of the prostate on MR images, to constrain the segmentation of the gland. The proposed method has been evaluated on the public dataset PROMISE2012. Experimental results show that the proposed method achieves a mean Dice similarity coefficient of 91.6\% as compared to the manual segmentation. The experimental results indicate that the proposed method based on anatomy knowledge can achieve satisfactory segmentation performance for prostate MRI.",

keywords = "Anatomy, Deep learning, Image segmentation, Location constraint, MRI, Prostate, Shape prior knowledge",

author = "Lei Tao and Ling Ma and Maoqiang Xie and Xiabi Liu and Zhiqiang Tian and Baowei Fei",

note = "Publisher Copyright: {\textcopyright} 2021 SPIE.; Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling ; Conference date: 15-02-2021 Through 19-02-2021",

year = "2021",

doi = "10.1117/12.2581893",

language = "English",

series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

publisher = "SPIE",

editor = "Linte, \{Cristian A.\} and Siewerdsen, \{Jeffrey H.\}",

booktitle = "Medical Imaging 2021",

address = "United States",

}

Tao, L, Ma, L, Xie, M, Liu, X, Tian, Z & Fei, B 2021, Automatic segmentation of the prostate on MR images based on anatomy and deep learning. 在 CA Linte & JH Siewerdsen (编辑), Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling., 115981N, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, 卷 11598, SPIE, Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling, Virtual, Online, 15/02/21. https://doi.org/10.1117/12.2581893

Automatic segmentation of the prostate on MR images based on anatomy and deep learning. / Tao, Lei; Ma, Ling; Xie, Maoqiang 等.
Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling. 编辑 / Cristian A. Linte; Jeffrey H. Siewerdsen. SPIE, 2021. 115981N (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; 卷 11598).

科研成果: 书/报告/会议事项章节 › 会议稿件 › 同行评审

TY - GEN

T1 - Automatic segmentation of the prostate on MR images based on anatomy and deep learning

AU - Tao, Lei

AU - Ma, Ling

AU - Xie, Maoqiang

AU - Liu, Xiabi

AU - Tian, Zhiqiang

AU - Fei, Baowei

PY - 2021

Y1 - 2021

N2 - Accurate segmentation of the prostate has many applications in the detection, diagnosis and treatment of prostate cancer. Automatic segmentation can be a challenging task because of the inhomogeneous intensity distributions on MR images. In this paper, we propose an automatic segmentation method for the prostate on MR images based on anatomy. We use the 3D U-Net guided by anatomy knowledge, including the location and shape prior knowledge of the prostate on MR images, to constrain the segmentation of the gland. The proposed method has been evaluated on the public dataset PROMISE2012. Experimental results show that the proposed method achieves a mean Dice similarity coefficient of 91.6% as compared to the manual segmentation. The experimental results indicate that the proposed method based on anatomy knowledge can achieve satisfactory segmentation performance for prostate MRI.

AB - Accurate segmentation of the prostate has many applications in the detection, diagnosis and treatment of prostate cancer. Automatic segmentation can be a challenging task because of the inhomogeneous intensity distributions on MR images. In this paper, we propose an automatic segmentation method for the prostate on MR images based on anatomy. We use the 3D U-Net guided by anatomy knowledge, including the location and shape prior knowledge of the prostate on MR images, to constrain the segmentation of the gland. The proposed method has been evaluated on the public dataset PROMISE2012. Experimental results show that the proposed method achieves a mean Dice similarity coefficient of 91.6% as compared to the manual segmentation. The experimental results indicate that the proposed method based on anatomy knowledge can achieve satisfactory segmentation performance for prostate MRI.

KW - Anatomy

KW - Deep learning

KW - Image segmentation

KW - Location constraint

KW - MRI

KW - Prostate

KW - Shape prior knowledge

UR - https://www.scopus.com/pages/publications/85105589470

U2 - 10.1117/12.2581893

DO - 10.1117/12.2581893

M3 - Conference contribution

AN - SCOPUS:85105589470

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2021

A2 - Linte, Cristian A.

A2 - Siewerdsen, Jeffrey H.

PB - SPIE

T2 - Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling

Y2 - 15 February 2021 through 19 February 2021

ER -

Automatic segmentation of the prostate on MR images based on anatomy and deep learning

摘要

出版系列

会议

联合国可持续发展目标

访问文件

其它文件与链接

指纹

引用此