Towards an Intelligent Framework for Personalized Simulation-enhanced Surgery Assistance: Linking a Simulation Ontology to a Reinforcement Learning Algorithm for Calibration of Numerical Simulations

Nicolai Schoch, Vincent Heuveline

Abstract


Evolving our previous research results in the context of cognition-guidance and patient-specifity for simulation-enhanced cardiac surgery assistance, in this work we further investigate on (1) a machine learning framework which allows to patient-individually calibrate soft tissue material parameters for subsequent simulation, and (2) a profound knowledge management framework which may enhance the ontology-driven overall setup of the cognition-guided surgery simulation in a clinic environment. Rather than being a closed research work with an in-depth theory backup and a complete evaluation, we here present a technical report and some interesting experimental works that are to serve for further research and development.

Keywords


Simulation Calibration; Parameter Identification; Personalized Numerical Simulation; Machine Learning; Reinforcement Learning; Knowledge Modeling; Simulation Ontology

Full Text:

PDF



DOI: https://doi.org/10.11588/emclpp.2017.05.42079

URN (PDF): http://nbn-resolving.de/urn:nbn:de:bsz:16-emclpp-420793