ORGANISATION/COMPANYUniversité Clermont Auvergne
RESEARCH FIELDComputer science
RESEARCHER PROFILEFirst Stage Researcher (R1)
APPLICATION DEADLINE28/06/2020 00:00 - Europe/Brussels
LOCATIONFrance › AUBIERE
TYPE OF CONTRACTTemporary
HOURS PER WEEK35 H
OFFER STARTING DATE01/10/2020
IS THE JOB RELATED TO STAFF POSITION WITHIN A RESEARCH INFRASTRUCTURE?Yes
Automatic segmentation of structures of the human deep brain from
nuclear magnetic resonance imaging
Supervisor: Omar Ait-Aider
Laboratory: Institut Pascal
Email and phone: firstname.lastname@example.org
Co-advisor(s): JJ. Lemaire, C. Teulière
Abstract (up to 10 lines):
The objective of this thesis is to allow the identification and automatic
localization of deep brain structures from 3D MRI imagery. The challenge is
to get as close as possible to the individual architecture, taking into account
intra-class variability for more precise clinical targeting. The difficulty in
obtaining large datasets pushes us to favor unsupervised methods. We will
thus consider of proposing an original method combining the anatomical
preconceptions with the machine learning methods to obtain a fine 3D
segmentation. Data fusion methods will also be considered in order to take
advantage of the complementarity of the different types of imagery.
• MS Degree or equivalent with major in Image processing, Machine
learning, or other related subjects
• Experience in medical imaging processing and 3D segmentation are
• Strong programming skills (C/C++, Python) and experience with deep
learning framework (TensorFlow/Keras/PyTorch…)
• Interest in medical applications and collaboration with clinicians
• Good spoken and written English
Keywords: 3D segmentation, Deep learning, MRI imaging, Deep brain, Data
Description (up to 1 page):
In brain MRI analysis, image segmentation is commonly used to measure and visualize the
anatomical structures of the brain, to analyze brain changes, to delineate pathological regions
and for surgical planning and image-guided interventions. In last decades, various
segmentation techniques of different precision and degree of complexity have been
developed and reported in the literature.
Recently, new methods using deep learning techniques for brain MRI segmentation have been
proposed [1-4]. They usually focus on well-known and identified structures of the brain.
The objective of the Ptolemee project, which brings together clinicians and researchers in
image analysis, is to map the deep and little-known regions of the brain. In this context, the
objective of this thesis is to develop an automatic segmentation method using a deep learning
approach based on a limited dataset of MRI images manually contoured by expert clinicians.
The dataset is necessarily limited because the manual contouring of these large and complex
MRI data is a tedious and difficult task for clinicians. The solution developed must therefore
rely on learning techniques while making the best use of more formal state-of-the-art models
in order to minimize the need for supervision.
First, the candidate will have to acquire the state of the art on brain MRI segmentation
methods. This step will be based on work currently in progress within the Institut Pascal.
In a second step the candidate will have to propose and implement original methods of
segmentation of the deep brain, taking into account the specificities of this dataset. Particular
emphasis will be placed on the combination of the a priori information provided by an
accurate atlas of the deep brain  and data-based deep learning approaches.
Finally, the contribution of the fusion of different imaging modalities for this segmentation
will be studied.
Références (up to ½ page):
•  Coupé et al., « AssemblyNet: A large ensemble of CNNs for 3D
Whole Brain MRI Segmentation », 2019
•  A. Guha Roy, S. Conjeti, N. Navab, and C. Wachinger, “QuickNAT: A
fully convolutional network for quick and accurate segmentation of
neuroanatomy,” NeuroImage, vol. 186, Feb. 2019.
•  A. de Brebisson and G. Montana, “Deep Neural Networks for
Anatomical Brain Segmentation,” in IEEE CVPR Workshops, 2015
•  C. Wachinger, M. Reuter, and T. Klein, “DeepNAT: Deep
convolutional neural network for segmenting neuroanatomy,”
NeuroImage, Apr. 2018.
•  Lemaire Jean-Jacques, De Salles Antonio, Coll Guillaume, El Ouadih
Youssef, Chaix Rémi, Coste Jérôme, Durif Franck, Makris Nikos,
Kikinis Ron, « MRI Atlas of the Human Deep Brain », in Frontiers in
Neurology vol.10, 2019
How to candidate?
Send CV, motivation letter and Master transcript to email@example.com
REQUIRED EDUCATION LEVELOther: Master Degree or equivalent
EURAXESS offer ID: 528414
The responsibility for the jobs published on this website, including the job description, lies entirely with the publishing institutions. The application is handled uniquely by the employer, who is also fully responsible for the recruitment and selection processes.
Please contact firstname.lastname@example.org if you wish to download all jobs in XML.