Aarhus University | Full Time | Aarhus, Denmark
Perhaps off field, but I'm recruiting a PhD student in Aarhus University in Denmark (in person or hybrid). It's salaried with pension contributions and help relocating, and the agreed pay is equivalent to a graduate job in Denmark.
Title: Multimodal Medical Imaging – AI and Quantum Physics for Diagnostics
Call for applications for a fully financed PhD fellowship
Project description:
Life creates order in a universe where disorder always increases. The set of chemical reactions enabling this —metabolism— changes as we eat, sleep, exercise, or become unwell. This PhD studentship is for a transformative research project integrating physics, mathematics, artificial intelligence (AI), and cutting-edge medical imaging to unravel and quantify these metabolic processes inside people.
Hyperpolarized MRI (HP-MRI) uses quantum mechanics to boost the sensitivity of MRI by orders of magnitude, allowing real-time imaging of metabolites like pyruvate as they are processed in vivo. Unlike conventional imaging techniques such as PET, HP-MRI avoids ionizing radiation, measures multiple metabolites simultaneously, and directly maps reactions like glycolysis and oxidative phosphorylation. It is uniquely powerful for diagnosing and monitoring diseases like cancer, multiple sclerosis, and Alzheimer’s, with its clinical potential currently being explored in multinational trials.
The EU-funded project "Quantum+AI for Diagnostics" (Q-AID) tackles one of the most fascinating challenges in medical imaging: mapping the relationship between underlying biological "truths" (e.g., tumor metabolism) and the signals captured by imaging systems, which are subject to noise, artefacts, and are imperfect. These mappings are deeply complex, involving layers of transformation —signal generation, reconstruction, and visualization— before even addressing the biological dynamics of living organisms. Metabolic systems are vast, interwoven reaction networks, where only a subset of processes can be measured directly. Using "physics-informed" neural networks, we aim to combine analytical knowledge with AI tools to infer unseen processes.
In this studentship, you will:
• Develop super-resolution models to overcome the spatial resolution limitations of HP-MRI, leveraging data from conventional imaging
• Model imaging processes mathematically, bridging biological truths and observable signals.
• Design AI approaches for semantic segmentation and metabolic classification, integrating multimodal imaging datasets.
• Quantitatively unify HP-MRI with other modalities like PET and CT.
• Work closely with clinicians to build life-saving tools that will impact patient care
See here to apply: https://phd.health.au.dk/application/opencalls/multimodal-me...