Using eye-tracking and correlates of brain activity to predict quality scores

Subjective quality tests give important information about the user's perception on the quality of an audio or video signal. These tests are helpful but usually lack information about the user's cognitive state. This might be crucial when consuming a service, as lower quality services may lead to a less frequent use. Therefore, physiological measures have become more popular in the domain of quality assessment as they give more detailed insights about e.g. fatigue and attention of a person. In the current study we used spatially degraded videos and obtained 1) standard subjective quality judgments, 2) eye movement parameters and 3) brain activity using electroencephalography (EEG). The results indicate more focused attention on the degraded area once participants found that area, as pupil diameter increases and the proportion of alpha activity decreases. On the basis of the data a model is proposed which predicts the quality scores from the obtained physiological measures.