A Real Time Human Emotion Recognition System Using Respiration Parameters and ECG

In the field of research on computer identification of emotion, physiological signals play an important role. The selection of the specific physiological input is dependent on its contribution to the emotion. In this research paper, light has been thrown on fusion of paramount physiological signals. The four types of physiological signals taken into account are: Electrocardiogram (ECG), Respiratory Rate (RR), Blood Pressure and Inhale-Exhale temperature of respiration. The research work done on this area is found to be minimal For first three signals, time domain features were extracted with a sensor system and an Intelligent processor. The system was trained using a feedback neural network and tested with unknown class inputs. To elicit emotion, short video sequences of 180 s are used. The videos were played in a laptop and kept at a distance of 1 m away from the subject under investigation. The results obtained are encouraging with the highest accuracy of 96.6% for happy and lowest of 70.38% for disgust with an average accuracy of 80.28%.