A stretch-sensing soft glove for interactive hand pose estimation

We present a stretch-sensing soft glove to interactively capture full hand poses with high accuracy and without requiring an external optical setup. Our device can be fabricated with simple tools available in most fabrication labs. The pose is reconstructed from a capacitive sensor array embedded in the glove. We propose a data representation that allows deep neural networks to exploit the spatial layout of the sensor itself. The network is trained only once, using an inexpensive off-the-shelf hand pose reconstruction system to gather the training data. The per-user calibration is then performed on-the-fly using only the glove.