The use of bio-signals for human computer interaction provides a good alternative to standard touch based interfaces, offering subjects with motor impairments an alternative means of communication. This research area investigates the use of electrooculographic (EOG) signals, measured via wireless EOG glasses, to interface with a real-time speller application.
Bio-signal based human computer interface (HCI) systems are a good alternative to standard touch-based interfaces, offering subjects with motor impairments an alternative means of communication. This work investigates the use of electrooculography (EOG) in such HCIs, where specifically the use of a wireless EOG glasses currently on the market, known as the JINS MEME, comprising only three dry electrodes, is compared to the standard, gel-based, six electrode configuration. A novel threshold-based classification algorithm was developed to classify saccades and blinks using EOG data from these two different eye movement recording modalities. Results show that a saccade classification accuracy of 64.57 per cent and a blink accuracy of 95.13 per cent were obtained using MEME acquired EOG data, which were found to be comparable to the 75.32 and 93.50 per cent saccade and blink classification accuracies obtained using the conventional setup. Such results confirm that the JINS MEME is a good alternative EOG recording modality, which could be used in HCI applications.
In this project, a novel, real-time speller application was also developed to compare the two EOG recording modalities in a real-time environment. This application only requires a training session which is 64 seconds or 80 seconds long using the MEME or the gel-based electrodes respectively, implying that it could be easily used in everyday life. Average writing speeds of 7.11 and 6.44 letters per minute were achieved when the speller was interfaced using the MEME and the gel-based electrodes, respectively. Such results compared well to the writing speed of 7.37 letters per minute obtained when the keyboard was interfaced by a camera-based eye gaze tracker.
Lead researchers:
