TenseInput: Augmenting Gesture Interaction with Voluntary Muscle Contraction GIX, University of Washington | Summer 2018

Gesture interaction can provide natural user experiences, yet it is often limited by the number of available gestures. TenseInput aims to solve this problem by utilizing voluntary hand or forearm muscle contraction as a supplementary channel for gesture input. Wrist-worn accelerometer and EMG sensors can recognize voluntary muscle contractions during complex gestures. A usability study with six common gesture interaction scenarios confirms that TenseInput is easy to learn and reduces input time. Please see the above video for more information.

Roles

  • Designed and implemented the sensing scheme
  • Trained CNN and RNN networks to detect voluntary muscle contractions
  • Developed three of the six interaction scenarios
  • Conducted a usability study to evaluate the practicality of the technique
  • Preparing for a spring submission to¬†ACM IMWUT as the first author