Emotion Detection Using Physiological Signals

Abstract

The emotion detection employing physiological signals is more unfailing as related to imaging and audio techniques. The biosignals such as ECG, GSR, used for emotion detection. The raw physiological signals contain distinct noises attributable to the power-line interference, motion of the electrodes, muscle artifacts, baseline wander. Hence pre-processing and filtering the signals makes the data compatible with the human-computer interaction systems. We have filtered and pre-processed the ECG, GSR signals of the AMIGOS database. Also eliminated the baseline drift using the baseline wandering pathfinding algorithm, extracted the peaks in the GSR signal, and interbeat measurements of the ECG signals. The various time-domain parameters, computed which can differentiate between the various dimensions of Russell's model. Human-computer interaction systems can use these methods for diverse applications.