Residual semantic segmentation of the prostate from magnetic resonance images

Md Sazzad Hossain, Andrew P. Paplinski, John M. Betts

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

Abstract

The diagnosis and treatment of prostate cancer requires the accurate segmentation of the prostate in Magnetic Resonance Images (MRI). Manual segmentation is currently the most accurate method of performing this task. However, this requires specialist knowledge, and is time consuming. To overcome these limitations, we demonstrate an automatic segmentation of the prostate region in MRI images using a VGG19-based fully convolutional neural network. This new network, VGG19RSeg, identifies a region of interest in the image using semantic segmentation, that is, a pixel-wise classification of the content of the input image. Although several studies have applied fully convolutional neural networks to medical image segmentation tasks, our study introduces two new forms of residual connections (remote and neighbouring) which increases the accuracy of segmentation over the basic architecture. Our results, using this new architecture, show that the proposed VGG19RSeg can achieve a mean Dice Similarity Coefficient of 94.57%, making it more accurate than comparable methods reported in the literature.

Original languageEnglish
Title of host publicationNeural Information Processing
Subtitle of host publication25th International Conference, ICONIP 2018 Siem Reap, Cambodia, December 13–16, 2018 Proceedings, Part VII
EditorsLong Cheng, Andrew Chi Sing Leung, Seiichi Ozawa
Place of PublicationCham Switzerland
PublisherSpringer
Pages510-521
Number of pages12
ISBN (Electronic)9783030042394
ISBN (Print)9783030042387
DOIs
Publication statusPublished - 2018
EventInternational Conference on Neural Information Processing 2018 - Siem Reap, Cambodia
Duration: 13 Dec 201816 Dec 2018
Conference number: 25th
https://conference.cs.cityu.edu.hk/iconip/

Publication series

NameLecture Notes in Computer Science
PublisherSpringer
Volume11307
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceInternational Conference on Neural Information Processing 2018
Abbreviated titleICONIP 2018
CountryCambodia
CitySiem Reap
Period13/12/1816/12/18
Internet address

Keywords

  • Deep convolutional neural networks
  • MRI images
  • Prostate
  • Semantic segmentation

Cite this

Hossain, M. S., Paplinski, A. P., & Betts, J. M. (2018). Residual semantic segmentation of the prostate from magnetic resonance images. In L. Cheng, A. C. S. Leung, & S. Ozawa (Eds.), Neural Information Processing: 25th International Conference, ICONIP 2018 Siem Reap, Cambodia, December 13–16, 2018 Proceedings, Part VII (pp. 510-521). (Lecture Notes in Computer Science ; Vol. 11307 ). Cham Switzerland: Springer. https://doi.org/10.1007/978-3-030-04239-4_46
Hossain, Md Sazzad ; Paplinski, Andrew P. ; Betts, John M. / Residual semantic segmentation of the prostate from magnetic resonance images. Neural Information Processing: 25th International Conference, ICONIP 2018 Siem Reap, Cambodia, December 13–16, 2018 Proceedings, Part VII. editor / Long Cheng ; Andrew Chi Sing Leung ; Seiichi Ozawa. Cham Switzerland : Springer, 2018. pp. 510-521 (Lecture Notes in Computer Science ).
@inproceedings{b1c77d9da0f9421094ab54bdc9e8f1ad,
title = "Residual semantic segmentation of the prostate from magnetic resonance images",
abstract = "The diagnosis and treatment of prostate cancer requires the accurate segmentation of the prostate in Magnetic Resonance Images (MRI). Manual segmentation is currently the most accurate method of performing this task. However, this requires specialist knowledge, and is time consuming. To overcome these limitations, we demonstrate an automatic segmentation of the prostate region in MRI images using a VGG19-based fully convolutional neural network. This new network, VGG19RSeg, identifies a region of interest in the image using semantic segmentation, that is, a pixel-wise classification of the content of the input image. Although several studies have applied fully convolutional neural networks to medical image segmentation tasks, our study introduces two new forms of residual connections (remote and neighbouring) which increases the accuracy of segmentation over the basic architecture. Our results, using this new architecture, show that the proposed VGG19RSeg can achieve a mean Dice Similarity Coefficient of 94.57{\%}, making it more accurate than comparable methods reported in the literature.",
keywords = "Deep convolutional neural networks, MRI images, Prostate, Semantic segmentation",
author = "Hossain, {Md Sazzad} and Paplinski, {Andrew P.} and Betts, {John M.}",
year = "2018",
doi = "10.1007/978-3-030-04239-4_46",
language = "English",
isbn = "9783030042387",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
pages = "510--521",
editor = "Long Cheng and Leung, {Andrew Chi Sing} and Seiichi Ozawa",
booktitle = "Neural Information Processing",

}

Hossain, MS, Paplinski, AP & Betts, JM 2018, Residual semantic segmentation of the prostate from magnetic resonance images. in L Cheng, ACS Leung & S Ozawa (eds), Neural Information Processing: 25th International Conference, ICONIP 2018 Siem Reap, Cambodia, December 13–16, 2018 Proceedings, Part VII. Lecture Notes in Computer Science , vol. 11307 , Springer, Cham Switzerland, pp. 510-521, International Conference on Neural Information Processing 2018, Siem Reap, Cambodia, 13/12/18. https://doi.org/10.1007/978-3-030-04239-4_46

Residual semantic segmentation of the prostate from magnetic resonance images. / Hossain, Md Sazzad; Paplinski, Andrew P.; Betts, John M.

Neural Information Processing: 25th International Conference, ICONIP 2018 Siem Reap, Cambodia, December 13–16, 2018 Proceedings, Part VII. ed. / Long Cheng; Andrew Chi Sing Leung; Seiichi Ozawa. Cham Switzerland : Springer, 2018. p. 510-521 (Lecture Notes in Computer Science ; Vol. 11307 ).

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

TY - GEN

T1 - Residual semantic segmentation of the prostate from magnetic resonance images

AU - Hossain, Md Sazzad

AU - Paplinski, Andrew P.

AU - Betts, John M.

PY - 2018

Y1 - 2018

N2 - The diagnosis and treatment of prostate cancer requires the accurate segmentation of the prostate in Magnetic Resonance Images (MRI). Manual segmentation is currently the most accurate method of performing this task. However, this requires specialist knowledge, and is time consuming. To overcome these limitations, we demonstrate an automatic segmentation of the prostate region in MRI images using a VGG19-based fully convolutional neural network. This new network, VGG19RSeg, identifies a region of interest in the image using semantic segmentation, that is, a pixel-wise classification of the content of the input image. Although several studies have applied fully convolutional neural networks to medical image segmentation tasks, our study introduces two new forms of residual connections (remote and neighbouring) which increases the accuracy of segmentation over the basic architecture. Our results, using this new architecture, show that the proposed VGG19RSeg can achieve a mean Dice Similarity Coefficient of 94.57%, making it more accurate than comparable methods reported in the literature.

AB - The diagnosis and treatment of prostate cancer requires the accurate segmentation of the prostate in Magnetic Resonance Images (MRI). Manual segmentation is currently the most accurate method of performing this task. However, this requires specialist knowledge, and is time consuming. To overcome these limitations, we demonstrate an automatic segmentation of the prostate region in MRI images using a VGG19-based fully convolutional neural network. This new network, VGG19RSeg, identifies a region of interest in the image using semantic segmentation, that is, a pixel-wise classification of the content of the input image. Although several studies have applied fully convolutional neural networks to medical image segmentation tasks, our study introduces two new forms of residual connections (remote and neighbouring) which increases the accuracy of segmentation over the basic architecture. Our results, using this new architecture, show that the proposed VGG19RSeg can achieve a mean Dice Similarity Coefficient of 94.57%, making it more accurate than comparable methods reported in the literature.

KW - Deep convolutional neural networks

KW - MRI images

KW - Prostate

KW - Semantic segmentation

UR - http://www.scopus.com/inward/record.url?scp=85059022877&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-04239-4_46

DO - 10.1007/978-3-030-04239-4_46

M3 - Conference Paper

SN - 9783030042387

T3 - Lecture Notes in Computer Science

SP - 510

EP - 521

BT - Neural Information Processing

A2 - Cheng, Long

A2 - Leung, Andrew Chi Sing

A2 - Ozawa, Seiichi

PB - Springer

CY - Cham Switzerland

ER -

Hossain MS, Paplinski AP, Betts JM. Residual semantic segmentation of the prostate from magnetic resonance images. In Cheng L, Leung ACS, Ozawa S, editors, Neural Information Processing: 25th International Conference, ICONIP 2018 Siem Reap, Cambodia, December 13–16, 2018 Proceedings, Part VII. Cham Switzerland: Springer. 2018. p. 510-521. (Lecture Notes in Computer Science ). https://doi.org/10.1007/978-3-030-04239-4_46