Zonal segmentation in transrectal ultrasound images of the prostate through deep learning

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

1 Citation (Scopus)

Abstract

Segmentation of both prostatic and zonal boundaries in transrectal ultrasound images is of great value in current clinical practice and for advancing techniques in computer-assisted diagnosis and inter-modality fusion. In this work, we propose a deep-learning approach to automatically segment the prostate and its main zones. In comparison with conventional methods, this method shows increased accuracy and Dice coefficients for full-prostate delineation. Moreover, the mean deviation between the annotated and predicted contours decreased substantially. Albeit the method still requires validation for different scanners and configurations, its real-time inference rate highlights the potential of this technique to be applied in clinical practice.
Original languageEnglish
Title of host publication2018 IEEE International Ultrasonics Symposium (IUS)
Place of PublicationPiscataway
PublisherInstitute of Electrical and Electronics Engineers
Number of pages4
ISBN (Electronic)978-1-5386-3425-7
ISBN (Print)978-1-5386-3426-4
DOIs
Publication statusPublished - 17 Dec 2018
Event2018 IEEE International Ultrasonics Symposium, IUS 2018 - Kobe, Japan
Duration: 22 Oct 201825 Oct 2018

Conference

Conference2018 IEEE International Ultrasonics Symposium, IUS 2018
Abbreviated titleIUS 2018
CountryJapan
CityKobe
Period22/10/1825/10/18

Keywords

  • Image segmentation
  • Image edge detection
  • Ultrasonic imaging
  • Magnetic resonance imaging
  • Medical diagnostic imaging
  • Biopsy
  • Prostate Cancer
  • Transrectal Ultrasound
  • (Auto-matic) Segmentation
  • Deep Learning

Fingerprint

Dive into the research topics of 'Zonal segmentation in transrectal ultrasound images of the prostate through deep learning'. Together they form a unique fingerprint.

Cite this