Brain Segmentation in Rodent MR-Images Using Convolutional Neural Networks

Björn Sigurdsson; Sune Darkner; Stefan Horst Sommer; Kristian Nygaard Mortensen; Simon Sanggaard; Serhii Kostrikov; Maiken Nedergaard

Brain Segmentation in Rodent MR-Images Using Convolutional Neural Networks

Björn Sigurdsson
, Sune Darkner
, Stefan Horst Sommer
, Kristian Nygaard Mortensen
, Simon Sanggaard
, Serhii Kostrikov
, Maiken Nedergaard

Technical University of Denmark

University of Copenhagen

Research output: Contribution to conference › Conference abstract for conference › Research

Abstract

This study compares two different methods for the task of brain segmentation in rodent MR-images, a convolutional neural network (CNN) and majority voting of a registration based atlas (RBA) , and how limited training data affect their performance. The CNN was implemented in Tensorflow. The RBA performs better on average when using a training set with fewer than 20 images but the CNN achieves a higher median dice-score with a training set of 19 images.

Original language	English
Publication date	2018
Publication status	Published - 2018
Event	Joint Annual Meeting ISMRM-ESMRMB 2018 - Paris Expo Porte de Versailles, Paris, France Duration: 16 Jun 2018 → 21 Jun 2018 https://www.ismrm.org/18m/

Conference

Conference	Joint Annual Meeting ISMRM-ESMRMB 2018
Location	Paris Expo Porte de Versailles
Country/Territory	France
City	Paris
Period	16/06/2018 → 21/06/2018
Internet address	https://www.ismrm.org/18m/

OpenUrl availability

Check availability in DTU Findit

Cite this

@conference{0962d07367fd44f79b170f3dd7d01074,

title = "Brain Segmentation in Rodent MR-Images Using Convolutional Neural Networks",

abstract = "This study compares two different methods for the task of brain segmentation in rodent MR-images, a convolutional neural network (CNN) and majority voting of a registration based atlas (RBA) , and how limited training data affect their performance. The CNN was implemented in Tensorflow. The RBA performs better on average when using a training set with fewer than 20 images but the CNN achieves a higher median dice-score with a training set of 19 images.",

author = "Bj{\"o}rn Sigurdsson and Sune Darkner and Sommer, \{Stefan Horst\} and Mortensen, \{Kristian Nygaard\} and Simon Sanggaard and Serhii Kostrikov and Maiken Nedergaard",

year = "2018",

language = "English",

note = "Joint Annual Meeting ISMRM-ESMRMB 2018 ; Conference date: 16-06-2018 Through 21-06-2018",

url = "https://www.ismrm.org/18m/",

}

TY - ABST

T1 - Brain Segmentation in Rodent MR-Images Using Convolutional Neural Networks

AU - Sigurdsson, Björn

AU - Darkner, Sune

AU - Sommer, Stefan Horst

AU - Mortensen, Kristian Nygaard

AU - Sanggaard, Simon

AU - Kostrikov, Serhii

AU - Nedergaard, Maiken

PY - 2018

Y1 - 2018

N2 - This study compares two different methods for the task of brain segmentation in rodent MR-images, a convolutional neural network (CNN) and majority voting of a registration based atlas (RBA) , and how limited training data affect their performance. The CNN was implemented in Tensorflow. The RBA performs better on average when using a training set with fewer than 20 images but the CNN achieves a higher median dice-score with a training set of 19 images.

AB - This study compares two different methods for the task of brain segmentation in rodent MR-images, a convolutional neural network (CNN) and majority voting of a registration based atlas (RBA) , and how limited training data affect their performance. The CNN was implemented in Tensorflow. The RBA performs better on average when using a training set with fewer than 20 images but the CNN achieves a higher median dice-score with a training set of 19 images.

M3 - Conference abstract for conference

T2 - Joint Annual Meeting ISMRM-ESMRMB 2018

Y2 - 16 June 2018 through 21 June 2018

ER -

Brain Segmentation in Rodent MR-Images Using Convolutional Neural Networks

Abstract

Conference

OpenUrl availability

Fingerprint

Cite this