Real-Time Control of a Video Game Using Eye Movements and Two Temporal EEG Sensors
Table 1
Comparison table of EOG- and video-based eye-tracking techniques.
Criteria
EOG electrodes
Video-based eye tracking
Intrusiveness
Intrusive with electrodes attached to the face (i.e., electrodes mounted on the skin around the eye).
Intrusive for cameras attached to glasses; nonintrusive for cameras mounted independently.
Complexity
(i) Electrodes number reduces the portability of the technique (many electrodes attached on the face). (ii) EOG is a simple and easy method to measure eye movements and it is still commonly used clinically for testing eye movements in patients.
(i) The algorithm complexity of the image processing system. (ii) Calibration is a crucial problem: head distance, head and pupil range of rotation with respect to sagittal plane of the body must be estimated (in some case manual corrections are still needed).
Influence of noise
Facial muscles (EMG signal) can be influenced on EOG signal.
(i) Light: big problem for image processing. (ii) Head movement: must keep your eyes open and in the vision field of the camera. (iii) Hard to use it in real-life application (outside environment).
Processing time
Fast: training or calibration phase needed.
Long: training or calibration phase needed; image processing takes much memory.
Classification accuracy
High, but related to visual angle, number of electrodes, and algorithm applied.
High, but related to head angle, user environment, and algorithm applied.