Welcome to the NEATBrainS15 website. Many researchers are working on either adult or neonatal brain segmentation. In this challenge workshop we exploited two challenges that were previously organized in neonatal (NeoBrainS12) and adult (MRBrainS13) MR brain segmentation. The main objective was to get an insight into the main differences and similarities between these two segmentation tasks. We challenged researchers to evaluate their automatic algorithms to segment gray matter, white matter and cerebrospinal fluid on both the NeoBrainS12 and MRBrainS13 data, and submit a workshop paper discussing the major issues encountered when applying their method to both types of data. In addition, we invited researchers to submit a workshop paper describing a novel algorithm in either one or both of these fields, to get an overview of the state of the art methods in both areas. During the workshop, we organized a panel discussion with experts in the field to discuss this topic, as well as the option of setting up a larger repository of data (and annotations) that can be used by researchers in the field.
This workshop was organized in association with ISBI 2015 in New York. If you have any questions, feel free to sent an e-mail to neatbrains15@isi.uu.nl.
Workshop
The NEATBrainS workshop was held on Thursday April 16th 2015 in the New York Marriott at Brooklyn Bridge in New York, USA. We are looking back at a great workshop.
It was interesting to see that the two methods submitted by the NEATBrainS participants that were initially developed for neonatal MR brain segmentation, are now the top 2 in the adult MR brain segmentation challenge.
And it was great to listen to the talks by Guido Gerig and Simon Warfield. We also got a lot of positive feedback on the panel discussion. The discussion slides can be downloaded here. A special thanks to the members of the panel for the opportunity to hear their thoughts on these matters.
Program
| 8:30 – 8:35 | Welcome and Opening – Adriënne Mendrik |
| 8:35 – 9:10 | Guido Gerig Invited talk “4D Processing and Analysis of Longitudinal Infant Imaging Reveals Spatiotemporal Pattern of Early Brain Growth“ |
| 9:10 – 9:45 | Simon K. Warfield Invited talk on Neonatal and Fetal Brain Segmentation and Validation |
| 9:45 – 10:00 | Pim Moeskops – Short talk “Evaluation of an Automatic Brain Segmentation Method Developed for Neonates on Adult MR Brain Images” |
| 10:00 – 10:15 | Li Wang – Short talk “Learning-based Multi-source Integration forNeonatal and Adult Brain Image Segmentation” |
| 10:15 – 10:30 | Coffee break |
| 10:30 – 11:30 | Panel discussion* part I: Methodology |
| 11:30 – 12:10 | Panel discussion* Part II: Data Sharing |
| 12:10 – 12:15 | Closing remarks – Ivana Išgum |
* The panel discussions were led by Daniel Rueckert. The following experts participated in the panel: Christian Barillot, Guido Gerig, Wiro Niessen, Julia Schnabel, and Simon K. Warfield.
Details
Motivation and objective
Brain tissue segmentation in MRI scans of adults is one of the classic image analysis problems dating back to 1985. Segmentation of brain tissue in MRI scans of neonates, on the other hand, is a relatively new area of research that is posed as a much more challenging image analysis problem. The main challenges in neonatal brain MR scans as opposed to adult scans are higher levels of noise, reduced contrast between tissue types, imaging at different stages during rapid brain development, white matter myelination, and increased partial volume effects. Some of the major challenges in adult MR brain segmentation are currently to develop robust and reproducible quantification methods, in the presence of noise, artefacts, pathology, and reduced contrast.
The main academic objective of the NEATBrainS15 workshop is to compare and discuss the performance of algorithms in segmenting gray matter, white matter and cerebrospinal fluid in adult versus neonatal brain MRI scans, and to discuss the similarities and differences between these two areas of research.
Challenge
Medical image analysis challenges provide a great platform for evaluating and comparing image processing algorithms for specific tasks. In this workshop we would like to exploit two challenges that we have previously organized. The first challenge is the NeoBrainS12 challenge on brain tissue segmentation in MR scans of neonates. The second challenge is the MRBrainS13 challenge on brain tissue segmentation in MR
scans of adults. Both challenges are open for new contributions. Researchers can download the training and test data, apply their segmentation algorithms and submit their results. For more information about these challenges, please visit the NeoBrainS12 and MRBrainS13 challenge websites.
We challenge researchers to apply their algorithms to both these challenges.
Organizers
Adriënne M. Mendrik, PhD
Image Sciences Institute
University Medical Center Utrecht
The Netherlands
Ivana Išgum, PhD
Image Sciences Institute
University Medical Center Utrecht
The Netherlands
Prof. Max A. Viergever
Image Sciences Institute
University Medical Center Utrecht
The Netherlands
Advisory Board
Christian Barillot (Universitaire de Beaulieu, France)
Guido Gerig (The University of Utah, USA)
Koen van Leemput (Harvard Medical School, USA)
Wiro Niessen (Erasmus MC – University Medical Center Rotterdam, the Netherlands)
Sebastien Ourselin (University College London, UK)
Daniel Rueckert (Imperial College London, UK)
Julia Schnabel (University of Oxford, UK)
Martin Styner (UNC Chapel Hill, USA)
Simon Warfield (Harvard Medical School, USA)
Ali Gholipour (Harvard Medical School, USA)