Research & Academic Work

My academic research focused on computer vision, medical imaging, and surgical simulation—laying the foundation for VirtaMed's mission to advance surgical training through realistic digital environments.

PhD Dissertation

Variable Anatomical Models for Surgical Training Simulators

Institution

Swiss Federal Institute of Technology (ETH) Zurich

Period

2001–2004

Abstract

After the emergence of minimally invasive surgery, new training paradigms had to be incorporated into the teaching curriculum for an appropriate learning of this non-intuitive technique. Virtual reality was proposed as a novel enabling technology for realistic training of such interventions which implicitly excludes any risks for patients. This dissertation addresses the computational complexity of anatomical models required for effective surgical simulation training. It presents a novel statistical method and computational tumor growth models to enhance training effectiveness, establishing collaborations with clinical partners to validate approaches through clinical studies.

Key Contributions

  • Developed statistical methods for variable anatomical models applicable to diverse patient anatomies
  • Created computational tumor growth models to improve training realism
  • Established clinical validation framework for surgical simulators
  • Supervised 3 master's theses and 6 semester projects during PhD research

Preview, Reference & Figures

Full dissertation available at the ETH Research Collection: View on ETH Research Collection →

Figure 3.3
Sample figure — Upper row: variation of the cervix-fundus angle for the mean shape. Lower row: variation of the fundal depth.
Figure 5
Sample figure — Real polyp (left) and view of an artificial polyp in the uterine cavity. The cavity has been inflated to allow for a better view, as is done during surgery.

Master's Thesis

Non-rigid Registration of Diffusion Tensor Images

Institution

Harvard Medical School & Brigham and Women's Hospital

Period

2000–2001

Research Focus

Developed novel algorithms for non-rigid registration of diffusion tensor imaging (DTI), an advanced MR imaging technology. The research focused on accurately aligning 3D medical images across different patients while preserving anatomical variability—a critical capability for surgical planning and training personalization.

Impact & Legacy

  • Algorithms and findings formed the basis for ongoing research in medical imaging
  • Methods incorporated into lectures and textbooks on medical imaging worldwide

Preview & Download

Academic Background

Master of Science

Electrical Engineering | ETH Zurich (1996–2001)

Extra-curricular coursework in marketing, business administration, economics, and supply chain management

Postdoctoral Research

Surgical Planning Laboratory, Boston (2005–2007)

Explored patient-specific training for minimally invasive brain surgeries

Computer Vision Laboratory, ETH Zurich (2004–2005)

System integration for hysteroscopy training simulator; prototype served as basis for VirtaMed

Publications & Collaborations

Published numerous peer-reviewed papers in computer vision, medical imaging, and surgical simulation, with research presented at leading international conferences and venues. Collaborated with clinical institutions including Harvard Medical School, Brigham and Women's Hospital, and ETH Zurich's computer vision and surgical planning laboratories.

This research continues to inform VirtaMed's development of evidence-based surgical training methodologies and digital twin technology for advancing surgical proficiency worldwide.

From 2017 to 2024 we partnered with ETH Zurich on a research project exploring model-driven medicine, focusing on digital twins and simulation. Find out more in this interview with Prof. Joachim Buhmann, go deep by reading this PhD thesis from Ivan Ovinikov, or get a glimpse of some exciting results in this post.

Since 2021 we are a leading partner in the Surgical Proficiency program together with leading hospitals and research institutions across Switzerland.