Surface model extraction from indentation curves of hyperelastic simulation for abnormality detection

Abstract

Manual palpation for the detection of anomalies is not possible through the small incisions of Robotic Minimally Invasive Surgery. The proposed novel approach allows robotic palpation by deforming the tissue surface with an indenter and analyzing the corresponding induced surface shape for indications of the abnormalities underneath. Three-dimensional hyperelastic finite element models were used to simulate the tool-tissue interaction of a hemispherical indenter pushing downwards onto the tissue surface. Curve fitting methods were employed to characterize the indentation curve of the deformed surface of either normal or abnormal tissue with an empirical equation. By analyzing these equations, we developed volume-based and gradient-based methods to investigate how the tumor position affects the surface deformation behavior of the tissue.The results of the simulations indicate that there are obvious differences in the surface deformation between healthy and diseased tissue, due to the higher stiffness of the tumor. A significant advantage of the proposed method is that it greatly broadens the detection area by providing estimates on the direction and distance of the tumor from the surrounding area of the indentation site, compared with previous studies only predicting the presence of a tumor in the contact area.