Calibrating Human Perception Threshold of Video Distortion Using EEG

Human perception threshold of video distortion is vital for visual quality assessment. Traditionally, calibrating the perception threshold of distortion relies on subjective test, which may suffer from strategy and bias of the human. In this paper, electroencephalography (EEG) is used as a novel psychophysiological method to evaluate the human perception of quantification-aroused video distortion. By the means of classification based on linear discriminant analysis (LDA), the separability of event-related potentials (ERPs) aroused by human perception to video quality change is measured by the area under curve (AUC) of the receiver operating characteristic (ROC). Relating the EEG signals to behavior data, the sigmoid-typed relationship between the perceptibility of distortion and separability of the P300 component is discovered. Based on this relationship, the perceptibility of distortion can be evaluated by EEG signals alone, which provides a potential neurally informed method for the calibration of perception threshold of video distortion.