This paper introduces three geometric features, from deformed shape of a soft tissue, which demonstrate good correlation with probing force and maximum local stress. Using FEM simulation, 2D and 3D model of an in vivo porcine liver was built for different probing tasks. Maximum deformation angle, maximum deformation depth and width of displacement constraint of the reconstructed shape of the deformed body were calculated. Two neural networks were trained from these features and the calculated interaction forces. The features are shown to have high potential to provide force estimation either for haptic devices or to assess the damage to the tissue in large deformations of up to 40%.