Multi-Scale Convolutional-Stack Aggregation for Robust White Matter Hyperintensities Segmentation

Hongwei Li, Jianguo Zhang (Lead / Corresponding author), Mark Muehlau, Jan Kirschke, Bjoern Menze

Research output: Chapter in Book/Report/Conference proceedingConference contribution

184 Downloads (Pure)


Segmentation of both large and small white matter hyperintensities/lesions in brain MR images is a challenging task which has drawn much attention in recent years. We propose a multi-scale aggregation model framework to deal with volume-varied lesions. Firstly, we present a specifically-designed network for small lesion segmentation called Stack-Net, in which multiple convolutional layers are ‘one-by-one’ connected, aiming to preserve rich local spatial information of small lesions before the sub-sampling layer. Secondly, we aggregate multi-scale Stack-Nets with different receptive fields to learn multi-scale contextual information of both large and small lesions. Our model is evaluated on recent MICCAI WMH Challenge Dataset and outperforms the state-of-the-art on lesion recall and lesion F1-score under 5-fold cross validation. It claimed the first place on the hidden test set after independent evaluation by the challenge organizer. In addition, we further test our pre-trained models on a Multiple Sclerosis lesion dataset with 30 subjects under cross-center evaluation. Results show that the aggregation model is effective in learning multi-scale spatial information.

Original languageEnglish
Title of host publicationBrainlesion
Subtitle of host publicationGlioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries
EditorsAlessandro Crimi, Spyridon Bakas, Hugo Kuijf, Farahani Keyvan, Mauricio Reyes, Theo van Walsum
Place of PublicationSwitzerland
PublisherSpringer Verlag
Number of pages9
ISBN (Electronic)9783030117238
ISBN (Print)9783030117221
Publication statusPublished - 26 Jan 2019
EventMICCAI 2018: 21st International Conference on Medical Image Computing & Computer Assisted Intervention - Granada Conference Centre, Granada, Spain
Duration: 16 Sept 201820 Sept 2018
Conference number: 21

Publication series

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


ConferenceMICCAI 2018
Internet address


  • Deep learning
  • White matter hyperintensities

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'Multi-Scale Convolutional-Stack Aggregation for Robust White Matter Hyperintensities Segmentation'. Together they form a unique fingerprint.

Cite this