DC5 PhD in Mathematical Imaging: Conditional Diffusion for Reconstruction and Cardiac MRI Analysis

DC5 PhD in Mathematical Imaging: Conditional Diffusion for Reconstruction and Cardiac MRI Analysis

Research cut-off: 6 November 2026
Assessment: Promising with material uncertainties

Quick Decision

DC5 is a mathematics-led doctoral project linking Markus Haltmeier’s Applied Mathematics group at the University of Innsbruck (UIBK) with Agnes Mayr’s cardiovascular MRI research environment at the Medical University of Innsbruck (MUI). It will develop and mathematically analyse conditional diffusion models for reconstruction, segmentation and image analysis. Myocardial-scar analysis in cardiac MRI is the motivating application. The current official application page sets the call deadline at 10 December 2026 and the official project page still lists DC5. The supplied listing carries an expired marker, but its explicit deadline agrees with the current call. The listing gives an offer-start date of 1 March 2027, whereas the official call expects employment after selection/admission in early February 2027; confirm the contract date. Apply · DC5 project

The strongest fit is an applicant who can demonstrate substantial mathematics—especially analysis, linear algebra and probability/stochastics—along with Python programming and interest in inverse problems, generative models and machine learning. Prior medical-imaging knowledge is not required by the project description. Recent local work on cardiac-MRI scar segmentation and uncertainty makes the application setting tangible, but it is adjacent evidence, not a guarantee of DC5’s data, architecture, work packages or outcomes. Confirm the legal employer, doctoral registration, dataset/compute access, secondments and pay/fees before accepting an offer.

Research Problem and Doctoral Work

What Is Unsolved and Why It Matters

Conditional diffusion models can use information beyond the image being reconstructed or analyzed. DC5 describes open challenges in how to integrate conditional information, such as clinical metadata, and how to maintain robustness across imaging protocols and patient populations. The project proposes to develop and mathematically analyse such models for image reconstruction, resolution enhancement, data completion, segmentation and analysis, using myocardial-scar analysis in cardiac MRI as a motivating application. The advertised scope is mathematical method research; it does not assert clinical deployment or patient benefit. The topic is tentative, and applicants are encouraged to develop a proposal within its scientific scope. DC5 project description

Work, Methods and Potential Research Angles

The input emphasizes conditional diffusion models, mathematical analysis, reconstruction and segmentation. A focused proposal could ask what conditioning signal is appropriate for a task, how conditioning affects the inverse problem or uncertainty, and how methods should be assessed when acquisition protocols or population characteristics vary. It could also clarify which output—reconstruction, resolution enhancement, missing-data completion or scar segmentation—would be used to test the mathematical ideas. These are research questions suggested by the advertised challenges; no fixed architecture, dataset or performance threshold is public.

Recent local work in cardiac-MRI segmentation and uncertainty provides adjacent context, not evidence of DC5’s methods or results; individual contributions and limits are assessed below.

Recent Team and Field Evidence

• Schwab et al. (2025) developed a fully automated LGE cardiac-MRI infarct-segmentation pipeline. It used retrospective data from one institution, with 144 training and 152 test examinations; the paper notes absent external validation and scanner/vendor limits. It is local scar-segmentation context, not a DC5 conditional diffusion result. Paper
• Schwab, Mayr and Haltmeier (2024) proposed an unsupervised deep Gaussian mixture segmentation method and demonstrated it on myocardial-infarct segmentation using multi-sequence MRI. It illustrates a local mathematical/learning approach but is not a diffusion model. Paper
• Schwab, Haltmeier and Mayr (2025) studied disagreement-driven uncertainty quantification for LGE cardiac MRI. This is relevant to DC5’s robustness motivation, but it does not establish the future project’s methods or deliverables. Paper
• Schwab et al. (2023) proposed a 2D/3D error-correcting network for LGE scar segmentation and evaluated it on the EMIDEC challenge dataset. It is a preprint on a conventional segmentation strategy, not conditional diffusion. Preprint
• Poskaite et al. (published online 2024) compared cardiac-MRI enhancement approaches for infarct-border discrimination in 110 STEMI patients. It provides MUI-group context for scar imaging; it is not a DC5 task or evidence of future clinical use. Paper
• Med-cDiff (2023) evaluates conditional diffusion in medical-image generation tasks including super-resolution, denoising and inpainting. It is external field evidence and does not establish DC5’s architecture or cardiac-MRI performance. Paper
• Peng et al. (2023) use a conditional diffusion probabilistic model to generate brain-MRI volumes from conditioning slices. This demonstrates a different MRI synthesis task and should not be conflated with myocardial-scar reconstruction. Paper
• Favero et al. (2025) explore class-conditional diffusion models for medical-image classification, including uncertainty and explainability analyses. This is a separate task and external evidence, not DC5 validation. Preprint

The MUI cardiovascular MRI group identifies cardiac MRI and infarct-related research as an established local area. Its environment supports the relevance of DC5’s motivating application, but public sources do not specify DC5 access to a scanner, clinical data, compute, or patient-facing work. Cardiovascular MRI group

School, Programme and Field Context

The listed supervisors are Markus Haltmeier of UIBK Mathematics and Agnes Mayr of MUI Radiology. UIBK’s Applied Mathematics Group describes work in inverse problems, signal and image processing, and machine learning, balancing mathematical analysis with algorithms. The MUI cardiovascular MRI group includes Mayr and describes collaborative research focused on cardiac-MRI measures and prognosis after ST-elevation myocardial infarction. This is a coherent mathematics/radiology setting, though exact supervision roles and the division of work are not public. Applied Mathematics Group · Haltmeier profile · MUI cardiovascular MRI

The stated enrolment route is the UIBK Doctoral Programme in Mathematics: full-time, six semesters/180 ECTS, English-taught, leading to a PhD. Admission requires a relevant diploma/master’s degree or equivalent and language documentation; UIBK provides a separate doctoral-registration process. Employment and academic matriculation are related but distinct. The programme also describes transferable training, thematic activities and cross-disciplinary supervision. IGNIDTE offers structured courses, seminars and secondments, but specific partners are tentative. Confirm the administrative employer and which office manages the mathematics PhD registration. Mathematics doctorate · Doctoral admission · IGNIDTE training

QS lists UIBK at #333 in the 2027 overall World University Rankings. That is not a mathematics, radiology or medical-diffusion subject ranking; UIBK notes that ranking systems have different methods. The university is located in Innsbruck, Austria. QS profile · UIBK ranking guidance · UIBK locations

Supervisor and Lab Assessment

Mathematics and Cardiac-MRI Environment

Haltmeier’s group works on inverse problems, signal and image processing and machine learning. Mayr’s group is part of the MUI cardiovascular MRI research environment, which reports a sustained cardiac-MRI research programme and collaboration with cardiology. Recent local papers by overlapping teams cover myocardial-scar segmentation, unsupervised MRI segmentation and uncertainty quantification. These publications are strong evidence of methodological and application-area activity; they do not establish that DC5 will use a specific image cohort, scanner, software or clinical workflow. Ask how the supervisors will divide mathematical theory, model implementation and cardiac-MRI evaluation, and what computational and data resources are approved for the project.

Funding, Application and Living Cost

Contract and Dates

The current official call offers a three-year, 40-hour-per-week doctoral employment contract. As of 2026, the minimum gross salary is €3,776.10 per month, paid 14 times per year, plus applicable allowances and benefits. Net salary is communicated at interview. The listing says 1 March 2027 as the offer-start date; the current official call expects employment after selection and admission in early February 2027. Confirm the exact date and contract terms in writing. First-call announcement · UIBK salary information

The current official deadline is 10 December 2026. The input deadline field agrees and adds 23:59 Europe/Vienna; the official page confirms the date, while the application portal should be checked for the exact cutoff time. DC5 applicants are asked to select the project and submit a two-page research proposal. The official selection criteria include academic qualifications, research experience, motivation/project fit, proposal, English/communication and complementary skills. Online interviews are expected 18–31 January 2027. Applicants using generative AI in application documents must disclose tools and contributions; non-English documents require certified English translations, and there is no application fee. Application page

Eligibility and Living Costs

Formal eligibility requires no prior doctoral degree, a relevant master’s or equivalent degree before employment, and no more than 12 months of living, working or studying in Austria during the 36 months before 10 December 2026. Applicants must not already be permanently employed by the host where the research training takes place. Applicants may apply before graduation if the final certificate is available before employment starts. The European Commission and REA explain the general MSCA framework; the current call controls this recruitment. European Commission MSCA guide · REA MSCA eligibility

UIBK estimates student costs at about €1,200/month (€450 accommodation, €350 food and €400 study/recreation). OeAD’s national student estimate is about €1,300/month, with accommodation €450–700, food around €300 and other study/personal costs around €500. These are student figures, not an employee budget; they do not establish take-home affordability, doctoral fees or ÖH contributions. UIBK stay guidance · OeAD living costs

Career Market Context

AMS lists 4,921 data-scientist postings in Austria in 2025 and a generally positive 2026–2028 outlook for that broad occupation. Statistics Austria reported 123,400 vacancies across sectors in Q2 2026. These are broad labour-market indicators, not forecasts for academic mathematics, cardiac MRI or conditional-diffusion research roles. Adjacent pathways could include applied mathematics, medical imaging, reconstruction, data science, machine-learning research and academic work, but none is guaranteed. AMS data-scientist barometer · Statistics Austria vacancies

Applicant Fit and Preparation

Formal Requirements and Preferred Background

The formal screens are a relevant master’s/equivalent before employment, no existing doctorate, MSCA mobility compliance and no permanent host employment. Separately, DC5 seeks someone who enjoys mathematics and can demonstrate solid foundations in analysis, linear algebra and probability/stochastics, plus Python programming. Interest in inverse problems, generative models and machine learning is preferred. Mathematics is the most direct background; physics or computer science can fit with strong documented mathematics. The listing explicitly says prior knowledge of medical imaging is not required. A purely application-oriented profile without evidence of mathematical depth is a weaker match.

Evidence to Prepare

Prepare a transcript/course list or thesis examples that show the stated mathematical foundations. Include concise evidence of Python programming or computational research. A two-page proposal can identify a conditional-diffusion question, explain why conditioning and robustness matter, and define a high-level way to evaluate it within the project’s scope. These are preparation suggestions rather than additional formal eligibility rules.

Next Actions and Questions to Verify

Before Applying

• Recheck that DC5 is selectable and verify the 10 December 2026 deadline and exact portal cutoff time.
• Calculate the MSCA 12-in-36-month mobility period and prepare degree/expected-completion evidence.
• Confirm the full upload checklist, certified translations/legalization, two-page proposal and AI-use disclosure.
• Ask how project selection and proposal scope should be represented in the application.

At Interview

• Ask what conditioning information and robustness questions are in scope and how mathematical analysis will be balanced with implementation/evaluation.
• Confirm the project’s data, computing, MRI-facility and collaboration access, plus the supervision split.
• Clarify the offer date conflict: 1 March 2027 in the input versus early February 2027 in the current official call.
• Confirm whether two secondments are guaranteed and ask about partners, timing and cost coverage.

Before Accepting an Offer

• Obtain written confirmation of employer, administrative unit, PhD registration, contract dates, gross/net pay, allowances and fees/ÖH obligations.
• Confirm approved data, compute and research support for DC5 and any supplementary coursework.

Evidence and Sources

Official call and project

• IGNIDTE current call and application rules
• IGNIDTE DC5 project page
• IGNIDTE training programme
• IGNIDTE first-call announcement

School, programme and environment

• UIBK Mathematics doctoral programme
• UIBK doctoral admission
• UIBK ranking guidance
• QS World University Rankings 2027
• UIBK student living-cost estimate
• OeAD living-cost estimate
• UIBK location information

Supervisors and research teams

• UIBK Applied Mathematics Group
• Markus Haltmeier official profile
• MUI cardiovascular MRI group
• MUI radiology research report 2024

Funding and career context

• UIBK scientific staff salary information
• European Commission MSCA Financial Guide
• REA MSCA eligibility
• Statistics Austria vacancies
• AMS JobBarometer: data scientists

Recent papers and field context

• Schwab et al. (2025), automated myocardial infarct segmentation
• Schwab et al. (2023), error-correcting scar segmentation
• Schwab, Mayr and Haltmeier (2024), deep Gaussian mixture segmentation
• Schwab, Haltmeier and Mayr (2025), LGE MRI uncertainty
• Poskaite et al. (2024), cardiac MRI scar border comparison
• Med-cDiff conditional medical-image diffusion (2023)
• Conditional diffusion probabilistic model for brain MRI (2023)
• Favero et al. (2025), conditional diffusion classifiers