Super-résolution en IRM appliquée à la reconstruction de cartographies quantitatives isotropes // MRI super-resolution applied to the reconstruction of isotropic quantitative maps
|
ABG-138864
ADUM-74473 |
Thesis topic | |
| 2026-05-01 | Public funding alone (i.e. government, region, European, international organization research grant) |
INSA Lyon
VILLEURBANNE Cedex - Auvergne-Rhône-Alpes - France
Super-résolution en IRM appliquée à la reconstruction de cartographies quantitatives isotropes // MRI super-resolution applied to the reconstruction of isotropic quantitative maps
- Computer science
super-résolution, imagerie par résonance magnétique, quantification, problème inverse, modélisation
super resolution, MRI, quantification , inverse problem, model
super resolution, MRI, quantification , inverse problem, model
Topic description
Les images obtenues par IRM ne fournissent généralement pas directement une information quantitative sur les tissus imagés. La collecte de données quantitatives constitue un champ de recherche actif visant à produire des données objectives et reproductibles. Certains paramètres quantitatifs, tels que les temps de relaxation ou la densité de proton, sont reliés à des propriétés physiopathologiques des tissus et peuvent être obtenus via des séquences spécifiques pour améliorer le diagnostic de certaines pathologies (infarctus, surcharge de fer ou sclérose en plaques [1]). Les séquences d'IRM quantitatives requièrent généralement des temps d'acquisition plus longs, dépendant notamment du nombre de dimensions nécessaires à l'ajustement du modèle sous-jacent. En pratique, cela impose des compromis entre durée d'acquisition, résolution spatiale et rapport signal-sur-bruit (SNR), souvent au détriment de la résolution ou de la qualité d'estimation. L'objectif principal de cette thèse consiste à mettre au point des techniques de reconstruction de cartographies quantitatives à haute-résolution à partir de plusieurs images acquises à basse-résolution. La technique permettrait d'améliorer de manière significative le compromis entre durée d'acquisition, qualité d'estimation et résolution spatiale.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
MRI images do not generally provide direct quantitative information about the tissues being imaged. The collection of quantitative data is an active field of research aimed at producing objective and reproducible data. Certain quantitative parameters, such as relaxation times or proton density, are linked to the pathophysiological properties of tissues and can be obtained via specific sequences to improve the diagnosis of certain conditions (heart attack, iron overload or multiple sclerosis [1]). Quantitative MRI sequences generally require longer acquisition times, depending in particular on the number of dimensions required to fit the underlying model. In practice, this necessitates trade-offs between scan time, spatial resolution and signal-to-noise ratio (SNR), often at the expense of resolution or estimation quality. The main objective of this thesis is to develop techniques for reconstructing high-resolution quantitative maps from multiple low-resolution images. The technique would significantly improve the trade-off between acquisition time, estimation quality and spatial resolution.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Début de la thèse : 01/10/2026
WEB : https://www.creatis.insa-lyon.fr/site/fr/recrutement/offre-de-these-super-resolution-en-irm-appliquee-la-reconstruction-de-cartographies
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
MRI images do not generally provide direct quantitative information about the tissues being imaged. The collection of quantitative data is an active field of research aimed at producing objective and reproducible data. Certain quantitative parameters, such as relaxation times or proton density, are linked to the pathophysiological properties of tissues and can be obtained via specific sequences to improve the diagnosis of certain conditions (heart attack, iron overload or multiple sclerosis [1]). Quantitative MRI sequences generally require longer acquisition times, depending in particular on the number of dimensions required to fit the underlying model. In practice, this necessitates trade-offs between scan time, spatial resolution and signal-to-noise ratio (SNR), often at the expense of resolution or estimation quality. The main objective of this thesis is to develop techniques for reconstructing high-resolution quantitative maps from multiple low-resolution images. The technique would significantly improve the trade-off between acquisition time, estimation quality and spatial resolution.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Début de la thèse : 01/10/2026
WEB : https://www.creatis.insa-lyon.fr/site/fr/recrutement/offre-de-these-super-resolution-en-irm-appliquee-la-reconstruction-de-cartographies
Funding category
Public funding alone (i.e. government, region, European, international organization research grant)
Funding further details
Concours pour un contrat doctoral
Presentation of host institution and host laboratory
INSA Lyon
Institution awarding doctoral degree
INSA Lyon
Graduate school
160 EEA - Electronique, Electrotechnique, Automatique de Lyon
Candidate's profile
La personne retenue aura une formation liée au traitement de l'image/du signal, ou aux mathématiques appliquées. Une expérience ou formation liée aux problèmes inverses, à l'optimisation ainsi qu'à l'apprentissage automatique est également fortement recommandée. Une formation ou expérience en imagerie médicale, en particulier en IRM sera appréciée.
La maîtrise du langage Python, de l'anglais (oral et écrit) est obligatoire.
La ou le candidat.e aura la volonté de travailler et de s'intégrer dans un groupe pluridisciplinaire et multiculturel, avec une volonté forte de développer des méthodes innovantes pour l'imagerie médicale et ses applications.
The successful candidate will have a background in image/signal processing or applied mathematics. Experience or training in inverse problems, optimisation and machine learning is also strongly recommended. Training or experience in medical imaging, particularly MRI, would be an advantage. Fluency in Python and English (spoken and written) is essential. The candidate should be keen to work and integrate into a multidisciplinary and multicultural team, with a strong desire to develop innovative methods for medical imaging and its applications. Translated with DeepL.com (free version)
The successful candidate will have a background in image/signal processing or applied mathematics. Experience or training in inverse problems, optimisation and machine learning is also strongly recommended. Training or experience in medical imaging, particularly MRI, would be an advantage. Fluency in Python and English (spoken and written) is essential. The candidate should be keen to work and integrate into a multidisciplinary and multicultural team, with a strong desire to develop innovative methods for medical imaging and its applications. Translated with DeepL.com (free version)
2026-05-20
Apply
Close
Vous avez déjà un compte ?
Nouvel utilisateur ?
Get ABG’s monthly newsletters including news, job offers, grants & fellowships and a selection of relevant events…
Discover our members
Medicen Paris Region
Tecknowmetrix
Servier
ONERA - The French Aerospace Lab
ADEME
Nantes Université
Généthon
Aérocentre, Pôle d'excellence régional
ASNR - Autorité de sûreté nucléaire et de radioprotection - Siège
ANRT
Nokia Bell Labs France
Institut Sup'biotech de Paris
Groupe AFNOR - Association française de normalisation
Laboratoire National de Métrologie et d'Essais - LNE
TotalEnergies
SUEZ
Ifremer
