Offre de stage Master 2, équipe SIAM : « Uncertainty-informed multimodal fusion pour la segmentation du thrombus et des lésions ischémiques en IRM »

TITLE : Uncertainty-informed multimodal fusion for segmentation of thrombus and ischemic lesions in MRI

Dates: from February to August 2026, 5-6 months

Partnership : IBISC, Centre Hospitalier Sud-Francilien, University Johns Hopkins

Subject

Develop an uncertainty-informed multimodal fusion approach for thrombus and ische- mic lesion segmentation on MRI in acute ischemic stroke. The project leverages SWAN/PHASE, DWI/ADC, and TOF-MRA to synthesize hypoperfusion-relevant information and improve cli- nical decision support [1–5].

Position of the problem

Stroke is the leading cause of acquired disability in adults and the second leading cause of death worldwide [6]. In ischemic stroke, a thrombus occludes a cerebral artery, depriving downstream tissue of oxygen. Treatment selection (thrombectomy or throm- bolysis) requires accurate clot localization and estimation of the ischemic area from hyperacute multimodal MRI [1–5]. However, signal variability, partial redundancy across modalities, and low-contrast/high-noise regions introduce segmentation uncertainty that limits clinical accuracy [7–9]. Prior cross-modal attention work achieved reliable thrombus detection (rate ≈ 0.97) but only Dice ≈ 0.65, underscoring the need to make uncertainty an explicit variable in fusion.

Internship objectives

A segmentation model based on cross-modal attention enabled re- liable thrombus detection (rate ≈ 0.97) but achieved a Dice ≈ 0.65, highlighting the need to address uncertain regions. The intern will help implement an uncertainty-aware multimodal fusion system to improve thrombus and ischemia segmentation.

— Uncertainty map generation. Compute voxel-wise entropy/Bayesian variance maps to identify high-uncertainty zones (e.g., thrombus borders) :

U_entropy(x) = − Σ p_c(x) log p_c(x),

c

leveraging Bayesian deep learning tools for calibrated uncertainty [7–9].

• Uncertainty-guided attentive fusion. Inject uncertainty maps as attention masks and focus inter-modal fusion where uncertainty is high so secondary modalities add comple- mentary evidence (compatible with modern biomedical backbones such as U-Net) [10].
• Localized diffusion regularization (“focused blurring”). Apply a Gaussian blur weighted by local uncertainty and train a diffusion model on blurred images to reinforce contextual learning near the thrombus :

I_blurred(x) = (I ∗ Gσ(x))(x),   σ(x) 𝖺 U (x),

using recent denoising diffusion frameworks [11].

• Backbones & estimators. Use 3D U-Net / Cross-Modal Attention Network / Diffusion Model ; estimate uncertainty via MC-dropout, deep ensembles, and Bayesian learning [7– 9] ; quantify cross-modal information with mutual information

I(M₁; M₂) = H(M₁) + H(M₂) − H(M₁, M₂),

and partial information decomposition (PID) to dissect redundancy, complementarity, and synergy [12, 13].

• Experimental comparison. Compare focused blur vs. global uniform blur ; assess Dice, sensitivity, and clinical precision (with modality-specific cues from SWI/PHASE, DWI/ADC, and TOF-MRA) [1–5].

Application & Expected Impact

• Datasets & Multimodal MRI from stroke patients (SWAN/PHASE, DWI/ADC, TOF-MRA) ; experiments on CHSF, MATAR, and ISLES2022. Environment : Py- Torch/Python, RTX 3090, CHSF–IBISC collaboration.
• Clinical utility. Improved thrombus and ischemia segmentation enables more reliable estimation of penumbral “mismatch,” supporting reperfusion triage when onset is un- certain and enhancing treatment benefit prediction.
• (i) Prototype of an uncertainty-guided multimodal segmentation model ;

(ii) quantitative analysis linking uncertainty to mutual information ; (iii) visual reports of clinically high-uncertainty regions ; (iv) manuscript targeting IEEE TMI or Frontiers in Neuroinformatics.

Expected results

• Prototype of an uncertainty-guided multimodal segmentation model ;
• Quantitative analysis of the link between uncertainty and mutual information ;
• Reporting and visualization of clinically high-uncertainty regions ;
• Preparation of a manuscript for IEEE Transactions on Medical Imaging or Frontiers in Neuroinformatics.

Candidate profile

We look for strongly motivated candidates (i) coming from a math, phy- sics, computer science or engineering diploma (ii) having a strong mathematical background, notably in linear algebra, analysis, probability and statistics, in machine learning and deep learning (iii) having good programming skills on some scientific language, preferably python.

Knowledge of medical imaging, particularly MRI, is not required, but is a strong plus. Know- ledge of basic optimization theory is also appreciated.

Practical information

The intern will be mainly hosted at the UFR science and technology (40 rue du Pelvoux) close to the city center. However, he/she may also spend some periods at the Hospital of Corbeil.

The monthly internship gratification is of about €670.

Application procedure : send a motivation letter, a CV and your University transcript (relevé de notes) since 1st year BSc to {vincent.vigneron,hichem.maaref}@univ-evry.fr.

What we offer

Hands-on experience with cutting-edge AI techniques for medical imaging.

Tackle real-world, high-impact healthcare problems using deep learning. Close mentorship from experienced researchers at the IBISC laboratory.

Opportunities to co-author publications and present your work at conferences. Continuation into PhD studies

Contact

{hichem.maaref, vincent.vigneron}@ibisc.univ-evry.fr,

Références

• E Mark Haacke, S Mittal, Z Wu, J Neelavalli, and Y-CN Cheng. Susceptibility weighted imaging (swi). Magnetic Resonance in Medicine, 52(3) :612–618,
• Àngel Rovira, Pilar Orellana, and José Alvarez-Sabín. The “susceptibility vessel sign” on t2*-weighted mri in acute ischemic stroke. Stroke, 40(2) :554–557, 2009.
• Steven Warach et Acute human stroke studied by diffusion-weighted mri. Annals of Neurology, 37(2) :231–241, 1995.
• DC Tong et Quantitative diffusion mri of acute ischemic stroke : Adc values predict tissue outcome. AJNR American Journal of Neuroradiology, 19(1) :104–110, 1998.
• Martin R Prince and Jeffrey 3d contrast in time-of-flight mr angiography. Journal of Magnetic Resonance Imaging, 12(5) :776–783, 2000.
• Valery L Feigin et al. Global, regional, and national burden of stroke and its risk factors, 1990–2019. The Lancet Neurology, 20(10) :795–820, 2021.
• Alex Kendall and Yarin What uncertainties do we need in bayesian deep learning for computer vision ? In NeurIPS, 2017.
• Yarin Gal and Zoubin Dropout as a bayesian approximation : Representing model uncertainty in deep learning. In ICML, pages 1050–1059, 2016.
• Balaji Lakshminarayanan, Alexander Pritzel, and Charles Simple and scalable predictive uncertainty estimation using deep ensembles. In NeurIPS, 2017.
• Olaf Ronneberger, Philipp Fischer, and Thomas U-net : Convolutional networks for biomedical image segmentation. In MICCAI, pages 234–241, 2015.
• Jonathan Ho, Ajay Jain, and Pieter Denoising diffusion probabilistic models. In

NeurIPS, 2020.

• Paul L Williams and Randall D Nonnegative decomposition of multivariate informa- tion. arXiv :1004.2515, 2010.
• Amer Makkeh, Dirk O Theis, and Raul Broja-2pid : A robust estimator for bivariate partial information decomposition. Entropy, 23(10) :1274, 2021.

• Date de l’appel : 12/11/2025
• Statut de l’appel : Non pourvu
• Contacts : Vincent VIGNERON (PR Univ. Évry, IBISC équipe SIAM), Hichem MAAREF (PR IUT Evry, IBISC équipe SIAM), Jean-Philippe CONGE , vincentDOTvigneronATuniv-evryDOTfr, hichemDOTmaarefATuniv-evryDOTfr,
• Sujet de stage niveau Master 2 (format PDF)
• Web équipe SIAM