Sapienza University of Rome, ECONA, University of Perugia, COGISEN
Explicit and Implicit Measures in Video Quality Assessment
Abstract: This work investigates the relation between subjective Video Quality Assessment (VQA) metrics and psychophysiological measures of human interaction assessment such as gaze tracking, electroencephalography and facial expression recognition. Subjective quality assessment is based on deliberate judgement attributions of perceived quality and processes that human perceivers are not consciously aware of. Traditional VQA methods ask participants to deliberately assign a quality score to videos in terms of the perceptual video quality. A methodology combining psychophysiological measures with traditional VQA methods is rarely used in the literature. This paper describes a model of video quality assessment which takes into account both explicit and implicit measures of subjective quality, by addressing two questions: (1) Do traditional video quality assessment methods correlate with unaware/implicit psychophysiological measures of quality perception assessment? (2) What can the main psychophysiological methods add to traditional video quality assessment? Findings show that (1) psychophysiological measures are able to measure differences of perceptual quality in compressed videos in terms of number of fixations and that (2) both VQA methods and psychophysiological assessment methods combined are able to provide additional information about cognitive and affective processes of attribution of the affective factors that underlie the attribution of quality.