A nonparametric modeling approach of soft tissue deformation by ANFIS

This paper presents a nonparametric modeling approach to soft tissue deformation utilizing an Adaptive Neural Fuzzy Inference System (ANFIS). The model is tested with real data. In order to obtain a consistent set of experimental data, a variable-velocity electro-mechanical platform applies single-point force to deform a soft tissue sample. A Motion Capture system obtains the position of twenty markers on the surface of the sample tissue. With applied force and position data of the central marker as inputs and the position of the remaining markers as outputs, an ANFIS system was designed and trained. The trained estimator is tested with experimental data under artificial noise conditions. The estimation of the position for a particular marker compared with the Motion Capture position data shows that the algorithm performs with less than 1% error.