[PDF] Post-doctorat en traitement dimages médicales

View - Download Post-doctorat en traitement dimages médicales

Previous PDF

Next PDF

Post-doctorat en traitement d"images m

´edicales

R ´eseaux de neurones convolutifs sur graphes de superpixels pour la segmentation d"images m

´edicalesBr.A.In.IMT Atlantique, Brest, France

1

D escriptif

Domaines de rechercheintelligence artificielle, traitement d"images bio-m´edicales, apprentissage profond, seg-

mentation s ´emantique, th´eorie des graphes, morphologie math´ematique EncadrantsPierre-Henri Conze et Nicolas Farrugia, IMT Atlantique

CollaborationsVincent Gripon, IMT Atlantique, Montreal Institute for Learning Algorithms (MILA), Montr´eal et

Gwenol

´e Quellec, Inserm

LieuIMT Atlantique, Brest, France

D ´ebutMars 2019, pour une dur´ee de 10 mois (renouvelable) 2

P rojetde r echerche

Le projet adresse la probl

´ematique de la segmentation d"images m´edicales par le biais de techniques issues de l"intelligence artificielle. Afin d"accro ˆıtre la fiabilit´e et la rapidit´e des outils d"aide au diagnostic, nous souhaitons

etendre les r´eseaux de neurones`a un support s"adaptant au contenu de l"image : les superpixels. Les superpixels

sont des primitives visuelles g

´en´er´ees par agr´egation de pixels voisins formant, une fois regroup´es, des r´egions

homog `enes et r´eguli`eres.

L"apprentissage profond appliqu

´e`a des graphes de superpixels repr´esente une perspective prometteuse qui n

´ecessite d"adapter les techniques existantes de traitement de signal sur graphes`a la segmentation s´emantique.

Ce projet vise

`a proposer des contributions m´ethodologiques`a l"interface entre plusieurs disciplines (intelligence

artificielle, traitement d"images, th ´eorie des graphes, morphologie math´ematique) et`a contribuer`a des cas con- crets d"applications en sant ´e dont notamment le segmentation de masses en mammographies pour le diagnostic du cancer du sein. 1

5 COMMENT CANDIDATER ?

3

E nvironnement

FinancementInstitut Carnot T´el´ecom et Soci´et´e Num´erique

Etablissement d"accueilIMT Atlantique est une grande´ecole d"ing´enieur g´en´eraliste et un centre de recherche

internationald ´ependantduminist`ereenchargedel"industrieetdunum´erique. IssuedelafusiondeT´el´ecom

Bretagne et des Mines de Nantes, c"est une

´ecole de l"Institut Mines-T´el´ecom. Le campus de Brest de l"IMT

Atlantique est id

´ealement situ´e en bord de mer [1] et b´en´eficie d"une tr`es bonne qualit´e de vie. 4

P rofilr echerch

´e

niveau de formation : doctorat (traitement d"images m´edicales, machine learning, computer science...)

notions avanc´ees en math´ematiques appliqu´ees, traitement d"images, machine learning et deep learning

fort int´erˆet pour le domaine de la sant´e autonomie, bonne aptitude`a communiquer 5

C ommentcandidater ?

Candidature

farrugia@imt-atlantique.fr) incluant:

CV avec liste des contributions scientifiques

2 articles scientifiques repr´esentatifs du travail pr´ec´edemment r´ealis´e

lettre(s) de recommandation lettre de motivation1 http://inovadys.com/29/telecom_bretagne/ 2 Postdoc position: Convolutional neural networks on superpixel graphs for medical image segmentationBr.A.In.IMT Atlantique, Brest, France 1

S ummary

GoalsExtend convolutional neural networks on superpixel graphs to perform medical image segmentation.

Research areasartificialintelligence,biomedicalimageprocessing,deeplearning,semanticsegmentation,graph

theory, mathematical morphology AdvisorsPierre-Henri Conze and Nicolas Farrugia, IMT Atlantique. and Gwenolé Quellec, Inserm.

LocationIMT Atlantique, Brest, France

Start dateMarch 2019, for a duration of 10 months (renewable) Deadline for applicationAccepting applications now, will remain open until filled 2

R esearchpr oject

Our project addresses medical image segmentation using techniques basedon artificial intelligence. To efficiently

assist clinicians, computer-aided diagnosis and follow-up tools require reliable methods to segment anatomical

and pathological structures. In order to increase the reliability and speed of these methods, we aim at extending

convolutional neural networks to support regions that adapt to the medical content: superpixels.

homogeneous and regular regions. Deep learning applied to superpixel graphs represents a promising perspec-

tive,requiringtheadaptationofexistinggraphsignalprocessingtechniquestosemanticsegmentation. Thegener-

alization of convolutional neural networks to graphs has been formely studied for image classification. A research

effort is needed to adapt these approaches to medical image segmentation based on superpixel graphs.

The purposes of this study is to propose methodological contributions at the interface between several disciplines

(artificial intelligence, image processing, graph theory, mathematical morphology) and to contribute to concrete

medical applications including massive diagnosis and follow-up of breast cancer using mammographies. 1

5 HOW TO APPLY

3

E nvironment

FundingThepresentpostdoctoralpositionisfundedbytheTélécom SociétéNumériqueCarnotInstitute(Carnot

TSN) which carries out scientific resourcing.

Host organizationIMT Atlantique is a public technological university focusing on the training of engineers at the

MSc. level and junior researchers at the PhD level. The Brest campus of IMT Atlantique is ideally located at

the sea front

1and benefits from a high quality of life.

4

C andidatepr ofile

PhD in machine learning, biomedical imaging or computer science, Interest in the fields of health and artificial intelligence,

Ability to communicate in English, intermediate level in French, fluent English for reading/writing scientific

articles,

Experience in deep learning.

5

H owto apply

Send an email topierre-henri.conze@imt-atlantique.frandnicolas.farrugia@imt-atlantique.frwith the following: a full curriculum vitae including a list of scientific contributions up to two representative scientific articles or conference papers recommendation letters or contacts from former teachers/advisors a cover letter stating your motivation and fit for this project1 http://inovadys.com/29/telecom_bretagne/ 2quotesdbs_dbs13.pdfusesText_19

[PDF] un produit mondialisé l automobile composition

[PDF] principe de non malfaisance

[PDF] principe de bienfaisance infirmier

[PDF] principe ethique definition

[PDF] principe de bienfaisance

[PDF] formation wordpress pdf gratuit

[PDF] principes éthiques fondamentaux universels

[PDF] les principes de léthique biomédicale

[PDF] wordpress tutoriel français pdf

[PDF] 4 principes bioéthique

[PDF] cours wordpress pdf 2016

[PDF] manuel wordpress 4 pdf

[PDF] la motivation dans lentreprise pdf

[PDF] jurisprudence internationale définition

[PDF] principe juridique droit