PhD position on Deep Learning for MRI-guided Radiotherapy

We are looking for a PhD student to enable MR-only radiotherapy. For details, please follow this link.


PhD position on deep learning in the analysis of cardiovascular images

This position is available within the Deep learning in Medical Image Analysis program funded by Dutch Technology Foundation. We are looking for a PhD student on Dynamic Deep Learning analysis in cardiac MR. For more information please mail


Scientific software engineer on deep learning analysis of cardiovascular images

This position is available within the Deep learning in Medical Image Analysis program funded by Dutch Technology Foundation. For more information please mail


MSc project: Functional MRI of Premature Infants

Functional MRI is an increasingly used tool in research and clinical settings. Besides for adults, the technique is used nowadays also for mapping brain functions in children. Several reports even indicate that fMRI is feasible for infants, including neonates and prematurely born babies. In most cases, fMRI in these groups will involve passive tasks during natural sleep or sedation. Because of the size of the infant’s brain, standard tools for, for example, brain segmentation into ROIs (to be able to compute fMRI activation in each of these ROIs) and normalization (necessary for averaging over subjects) are not adequate and analysis of infant fMRI data will require dedicated tools.
We are looking for an MSc student in (biomedical) engineering, artificial intelligence, computer science, or comparable training, with an interest in medical image analysis and deep learning to design tools for this analysis.

MSc project: Automatic classification of spine injuries/fractures

Traumatic spinal injuries are a global public health concern in terms of care and costs. The incidence of spinal injuries ranges from 4.6% to 24.0%. Overall, about 50% of thoracolumbar injuries are unstable and can result in significant disability, deformity, and neurological deficit. Imaging and classification of the thoracolumbar fractures are an important part in the management of spinal injuries and for deciding the optimal treatment.

The AO Spine classification group has developed a new classification system based on the severity of injury. This differentiates three main types of injury: Type A lesions are compression lesions to the anterior column; Type B lesions are lesions of either the anterior or the posterior tension band; Type C lesions are dislocations / displacements.


Classification of spine fracture patterns into these complex schemes can be a time-consuming task, resulting in additional work for the surgeon/radiologist and even worse, failure in correctly diagnosing the fracture. Research has been conducted for computerized assessment of compression fractures through the detection of vertebral body height loss on midline sagittal sections of lumbar computed tomographic (CT) images and on three-dimensional volumetric renderings.

Segmentation and analysis of the spine on CT images, with direct quantitative assessment of bone and fractures of vertebral bodies, is of clinical importance for spine injury diagnosis. Appropriate classification in this context is the focus of this project.

You will tackle this problem by starting with the type A fractures, the compression fracture resulting in height loss of the vertebral body. You will develop and validate a (semi) automated tool using machine learning/pattern recognition for the detection and anatomic localization of vertebral body fractures on CT images.

We are looking for a motivated MSc student with a background in Biomedical Imaging/Engineering, Computer Science, Mathematics or similar. Good programming skills (preferably Python, C++) and an interest in machine learning (specifically deep learning) are required. If you are interested in this project, please contact dr. Ivana Išgum.