Visual Analytics for Intra-Operative Brain Shift Compensation in Ultrasound Data



During neurosurgery, Image-Guided Neurosurgical Systems (IGNSs) provide a rigid patient-to-image mapping that relates the pre-operative image data to an intra-operative patient coordinate system, allowing surgeons to achieve a radical tumor resection while avoiding damage to surrounding functioning brain tissue. However, intra-operative deformation of the brain, also known as brain shift, invalidates this mapping.

In order to increase the accuracy of neurosurgeries, an active registration framework is developed to compensate the brain shift [1]. However, the employed registration strategies are often complex, employing Machine Learning methods and carrying several sources of uncertainty, such as from different image modalities, deformation fields, or interactive adjustments of the doctors. The field of Visual Analytics can provide insight into this kind of data and processes, but registration visualization has not been vastly researched - with just a few examples in the literature [2,3]. 

[1] Luo, J., Toews, M., Machado, I., Frisken, S., Zhang, M., Preiswerk, F., Sedghi, A.,
Ding, H., Pieper, S., Golland, P., Golby, A., Sugiyama, M., & Wells III, W. M.: A featuredriven
active framework for ultrasound-based brain shift compensation. MICCAI’18,
Granada, Spain.

[2] Smit, N.N., Haneveld, B.K., Staring, M., Eisemann, E., Botha, C.P. and Vilanova, A., 2014, September. RegistrationShop: An Interactive 3D Medical Volume Registration System. In VCBM (pp. 145-153).

[3] Schlachter, M., Fechter, T., Jurisic, M., Schimek-Jasch, T., Oehlke, O., Adebahr, S., Birkfellner, W., Nestle, U. and Bühler, K., 2016. Visualization of Deformable Image Registration Quality Using Local Image Dissimilarity. IEEE transactions on medical imaging, 35(10), pp.2319-2328.


We would like to investigate novel approaches from the field of Visual Analytics to: 1) facilitate understanding in the registration approaches utilized for brain shift compensation, and 2) visualize, explore and analyze the uncertainty that accompanies these registration processes.

This project is offered as a Bachelor or a Master Thesis: depending on this, the goals of the thesis will be fine-tuned to match the level of studies.


  • Knowledge of English language (source code comments and the final report should be in English)
  • Knowledge of web-based APIs, like JavaScript D3 is advantageous
  • Knowledge of Python is advantageous
  • Knowledge of 3DSlicer ( is advantageous, but not necessary
  • Commitment to collaborate with international researchers and surgeons (primarily from
    Brigham and Women’s Hospital, Harvard Medical School)


The project should be implemented as a web-based application (to be discussed).


For more information please contact Jie Luo or Hsiang-Yun Wu ( or Renata Raidou (