Perceptual video quality evaluation by means of physiological signals

Abstract

The proliferation of multimedia technology and its wide adoption by users has created the need for more effective metrics for Quality of Experience (QoE). Objective video quality metrics usually under-perform in terms of perceptual quality, thus evaluation is usually performed offline by people, an arduous and time consuming task that is also affected by external conditions and by user preferences. The use of physiological signals, recorded from users exposed to multimedia stimuli, has the potential to offer a more robust and unbiased method for evaluating perceptual quality. In this work, we propose the evaluation of the perceptual quality of video by means of cerebral (Electroencephalography -EEG) and peripheral (Electrocardiography -ECG and Electromyography -EMG) physiological signals. A machine learning approach is employed in order to map features extracted from these signals to a subjective video quality scale. Five 4K video sequences were encoded at different quality levels using the state-of-the-art HEVC codec and their quality was evaluated by real users while recording their physiological signals. The quality levels decided by the proposed model were then evaluated against the user-provided MOSs and the results demonstrated the potential of the proposed method for accurate perceptual video quality evaluation.

Visual Interface Evaluation for Wearables Datasets: Predicting the Subjective Augmented Vision Image QoE and QoS

“…For more practical considerations, commercially-available consumer-grade hardware can be employed in the experimentation and data gathering. However, other physiological signals could be employed, as well, as examples for skin conductivity or heart rate show promise [34].…”

A common framework for the evaluation of psychophysiological visual quality assessment

“…Currently it is not clear if a psychophysiological response might be influenced, e.g., biased by the source content. Source content should therefore be chosen spanning a wide range of temporal and spatial information and to have neutral impact on the subjects [23]. A precise description of the stimuli utilized is thus crucial.…”

Affect Detection for Human-Horse Interaction

“…EEG was recorded at a sampling rate of 128 Hz using the Emotiv EPOC+ wireless EEG headset [26] that has 16 goldplated contact sensors that are fixed to flexible plastic arms of the Emotiv EPOC wireless headset and are placed against the head in locations aligned with the following locations according to the International 10-20 system: AF3, F7, F3, FC5, T7, P7, O1, O2, P8, T8, FC6, F4, F8, AF4, M1 and M2. ECG and EMG were recorded at a sampling rate of 256 Hz using two wireless Shimmer sensors [27], and using four standard electrodes placed on both lower ribs and clavicle for [28]. All devices were connected to a Raspberry Pi portable mini-computer that was powered by a standard off-the-shelf USB power bank, and the acquired signals were recorded and stored on the Raspberry Pi's micro SD memory card.…”