Efficient patient modeling for visuo-haptic VR simulation using a generic patient atlas

Andre Mastmeyer*, Dirk Fortmeier, Heinz Handels

*Corresponding author for this work
4 Citations (Scopus)


Background and Objective: This work presents a new time-saving virtual patient modeling system by way of example for an existing visuo-haptic training and planning virtual reality (VR) system for percutaneous transhepatic cholangio-drainage (PTCD). Methods: Our modeling process is based on a generic patient atlas to start with. It is defined by organ-specific optimized models, method modules and parameters, i.e. mainly individual segmentation masks, transfer functions to fill the gaps between the masks and intensity image data. In this contribution, we show how generic patient atlases can be generalized to new patient data. The methodology consists of patient-specific, locally-adaptive transfer functions and dedicated modeling methods such as multi-atlas segmentation, vessel filtering and spline-modeling. Results: Our full image volume segmentation algorithm yields median DICE coefficients of 0.98, 0.93, 0.82, 0.74, 0.51 and 0.48 regarding soft-tissue, liver, bone, skin, blood and bile vessels for ten test patients and three selected reference patients. Compared to standard slice-wise manual contouring time saving is remarkable. Conclusions: Our segmentation process shows out efficiency and robustness for upper abdominal puncture simulation systems. This marks a significant step toward establishing patient-specific training and hands-on planning systems in a clinical environment.

Original languageEnglish
JournalComputer Methods and Programs in Biomedicine
Pages (from-to)161-175
Number of pages15
Publication statusPublished - 01.08.2016


Dive into the research topics of 'Efficient patient modeling for visuo-haptic VR simulation using a generic patient atlas'. Together they form a unique fingerprint.

Cite this