| 摘要 |
For detecting affective events in a video, affective events being fast changes in a subject's emotional state while watching the video, a device is described, the device comprising receiving means INP for receiving samples of physiological signals from the subject while watching the video, the physiological signals comprising at least facial surface ElectroMyoGrams EMG and galvanic skin response GSR; and processing means PRC adapted for extracting feature vectors from buffered samples; wherein the processing means PRC is further adapted for using a supervised classifier GPC for determining and outputting whether the extracted feature vectors is indicative of an affective event, the supervised classifier GPC being designed on basis of a database DB of labeled exemplary feature vectors. |
