3D Deep Learning for Dental Diagnostics

• Develop deep learning models for 3D intraoral scans, starting from point cloud models such as PointNet++ and DGCNN
• Extend these toward graph neural networks that operate directly on the mesh structure of the scans
• Use explainability methods to show which geometric features drive each prediction
• Validate the models with dentists and present your results at international conferences

If you have worked in 2D computer vision and want to move to 3D data, or have used graph neural networks in other domains, this project lets you apply these skills to clinical data. No prior experience with 3D data or dentistry is required.

• Master’s degree in computer science, data science, mathematics, physics, biomedical engineering or a related field
• Solid experience with deep learning, Python and PyTorch or a similar framework
• Preferred, experience with graph neural networks, point clouds or meshes, PyTorch Geometric or explainable AI, Research Field Computer science

Education Level Master Degree or equivalent, Master Degree or equivalent

Research Field Mathematics

Education Level Master Degree or equivalent

Research Field Physics

Education Level Master Degree or equivalent

What we offer

• PhD position at the Medical University of Vienna, 30 hours per week, 36 months
• Salary according to the collective agreement for Austrian universities (B1), currently EUR 2,832 gross per month, paid 14 times per year
• A group of eight PhD students and a postdoc at the interface of machine learning and dentistry, based inside the dental clinic
• Access to the MedUni Vienna high-performance computing cluster and travel funding for international conferences