3D Touch Force Estimation from Capacitive Images

DOI

Force plays a significant role in our daily interactions with the surrounding environment. However, most touchscreen interactions utilize only touch location, neglecting the potential for incorporating touch force as an additional input modality due to the difficulty in accurately estimating touch force without extra sensors. In this study, we propose a method for estimating three-dimensional relative touch forces, including pressure perpendicular to the touchscreen surface and shear force parallel to the surface, using two raw capacitive images. We collected two datasets comprising raw capacitive data from a touchscreen and corresponding forces measured by a triaxial force sensor. In the first dataset, participants performed press actions to apply pressure, while in the second dataset, they performed push actions to apply shear force to the touchscreen. Empirical experiments demonstrated that our proposed method outperformed existing force estimation methods, achieving mean absolute errors of 0.41 N, 0.44 N, and 0.40 N in the lateral, longitudinal, and vertical directions, respectively. Additionally, we conducted a user study with four tasks to assess the performance of our method in real-world scenarios, encompassing force gestures, pressure control, shear force control, and object manipulation using a combination of conventional touch input and force input. Comparisons with other methods demonstrated the superiority of our approach.