ESR 11: Surgical Planning through Hands-on Medical Image Editing

Early Stage Researcher  Zhongyun He
Host institution: Universidad Rey Juan Carlos (URJC, ES)
Supervisor: Prof. Miguel Otaduy, URJC
Co-supervisor: Carlos Illana, Advanced Healthcare Technology Product Manager, GMV (ES)
Clinical expertise: David Macías Verde, Medical Physicist, Hospital Universitario de Gran Canaria Dr. Negrín
Further institution involved: Université du Luxembourg (UL, LU)

Objectives

Given a medical image of the anatomical region of interest, the first objective will consist of setting up, in an automated manner, an efficient yet reliable deformation model. This will be achieved through a novel approach that will combine segmentationfree meshing with ground-breaking nonlinear homogenization methods. Inclusion of support for topological changes in the homogenized deformation model. Finally, a third objective will consist of the design of novel natural interaction techniques for the manipulation and deformation of medical images. The target scenario will take the form of an intuitive clinical design application executed on a touch screen (e.g., a tablet PC), where the clinical expert will manipulate the anatomical image directly.. Novel interaction methods will enable the mapping of natural 2D gestures to the manipulation and deformation of the underlying 3D anatomy. The intuitive planning methods will be tested on an intraoperative radiotherapy planning tool.

Expected Results

Surgical planning tools that allow direct tangible interaction with medical image data on touch screens, built on top of novel deformation models based on nonlinear homogenization and natural interaction techniques for the manipulation and deformation of medical images.

Planned research stays

  • URJC (5 months): Initial training on deformation and interaction methods.
  • GMV (4 months): Secondment for integration of natural interaction techniques for touch screens, and validation on clinical dataset.
  • URJC (8 months): Design and development of deformation models for surgical planning, based on nonlinear homogenization with support for topological changes.
  • UL (3 months): secondment for design of data-driven medical image deformation models
  • URJC (9 months): Design and development of deformation models for surgical planning, based on nonlinear homogenization with support for topological changes
  • GMV (3 months): Secondment for integration of natural interaction techniques for touch screens, and final validation on clinical datasets.
  • URJC (4 months): Thesis writing.

Further information of the project at URJC lab webdescription.