Text size
  • Small
  • Medium
  • Large
  • Standard
  • Blue text on blue
  • High contrast (Yellow text on black)
  • Blue text on beige

    Comparative Performance Analysis of a Commercial Wearable EOG Glasses for an Asynchronous Virtual Keyboard

    HCI 2018

    Proceedings of the 32nd International BCS Human Computer Interaction Conference (HCI 2018)

    Belfast, UK, 4 - 6 July 2018


    Nathaniel Barbara, Tracey A. Camilleri & Kenneth P. Camilleri



    Conventional electrooculography (EOG) signal acquisition systems, apart from requiring expensive equipment, necessitate electrode gel for conductivity and electrode wires, which may obstruct the field of view and restrict user movements, thus making the setup itself impractical for regular use. Recently however, a cheaper, wireless, gel-free, sleek and wearable alternative EOG recording device, inconspicuously incorporating dry electrodes mounted on an ordinary-looking pair of glasses, has been made commercially available. This work compares this device, known as the JINS MEME EOG glasses, against a gold-standard conventional gel-based EOG setup.

    Specifically, while gaze displacement estimation errors of 1.32±0.26° (and 1.67±0.26°) in the horizontal (and vertical) directions were obtained using the conventional setup, the corresponding errors obtained using the JINS MEME were 1.97±0.34° (and 1.85±0.30°). Saccades and blinks were also found to be reliably detected and labelled using these two EOG recording modalities, with average labelling accuracies exceeding 99%. The two modalities were also compared when used in real-time to interface with an asynchronous EOG-based virtual keyboard having a QWERTY layout. Specifically, average writing speeds across subjects of 11.9±4.4 and 9.9±3.6 characters per minute were obtained using the conventional EOG setup and the JINS MEME respectively, which have been shown to improve substantially with user experience with the system.

    These results have demonstrated that the JINS MEME offers a feasible wearable alternative to the conventional EOG setup, which is more practical for eye movement-based applications.


    PDF filePDF Version of this Paper (1,140kb)