Contribution
Comparison of Electrooculography (EOG) and camera-based eye blink detection
* Presenting author
Abstract:
Effective everyday communication involves not only verbal language, but also the nonverbal cues exchanged between conversational partners. A key challenge is recognizing when a conversation partner is taking turn, especially in complex situations with multiple interlocutors and distracting environments. Such nonverbal backchannel information can be expressed, among other measures, through eye blinking behavior. Therefore, eye blink synchronization across conversational partners could be a good predictor of communication success. In order to be able to detect and analyze eye blinks in real time in communication experiments, two methods have been compared in this work. In particular, a physiological measurement - Electrooculography (EOG) - was used and compared to an optical measurement approach. The results show that the EOG-based approach provides more information about the duration, amplitude and slope of eye blinks than the camera-based approach. In addition, a shorter latency can be obtained with EOG. Therefore, we conclude that EOG-based eye blink detection is a suitable approach to investigate the role of eye blinks in communication experiments with and without hearing devices.