Neural network modelling of soft tissue deformation for surgical simulation

Highlights：

• A new neural network methodology for modelling of soft tissue deformation for real-time, realistic, and stable surgical simulation.

• Two CNN models are developed to carry out soft tissue deformation via (i) neural propagation and (ii) dynamics by combining bioelectric energy propagation of soft tissues with mechanical deformation dynamics.

• The proposed methodology not only satisfies the real-time requirement of surgical simulation but also achieves the physical behaviours of soft tissue deformation. Further, it also achieves stable model dynamics for soft tissue simulation, but with similar computational efficiency to the explicit time integration.

• Simulation and experimental results demonstrate that the proposed method exhibits nonlinear force-displacement relationship and the associated nonlinear deformation behaviours of soft tissues. Further, the proposed method can accommodate not only homogeneous but also anisotropic and heterogeneous materials by simple modification of electrical conductivity values of mass points.

摘要

•A new neural network methodology for modelling of soft tissue deformation for real-time, realistic, and stable surgical simulation.•Two CNN models are developed to carry out soft tissue deformation via (i) neural propagation and (ii) dynamics by combining bioelectric energy propagation of soft tissues with mechanical deformation dynamics.•The proposed methodology not only satisfies the real-time requirement of surgical simulation but also achieves the physical behaviours of soft tissue deformation. Further, it also achieves stable model dynamics for soft tissue simulation, but with similar computational efficiency to the explicit time integration.•Simulation and experimental results demonstrate that the proposed method exhibits nonlinear force-displacement relationship and the associated nonlinear deformation behaviours of soft tissues. Further, the proposed method can accommodate not only homogeneous but also anisotropic and heterogeneous materials by simple modification of electrical conductivity values of mass points.