3D force prediction using fingernail imaging with automated calibration

Abstract

This paper demonstrates a system for 3D force prediction using fingernail imaging, in which video images of the human fingernail are used to predict the normal and shear forces that occur when the finger is in contact with a flat surface. The automated calibration uses a magnetic levitation haptic device (MLHD) whose flotor has been modified to apply forces to the human fingerpad. The system accurately predicts forces with an RMS error of 0.3N normal force, 6% of the full range of 10N, and a shear force error of 0.3N, 3% of the full range of ±2.5N. This paper also demonstrates that varying the number of pixels used to represent the finger between 100 and 500 pixels has little effect on the results, indicating that a real-time application could use low-resolution images without loss of accuracy.