International Journal of Scientific Engineering and Research (IJSER) ISSN (Online): 2347-3878 Index Copernicus Value (2015): 62.86 | Impact Factor (2015): 3.791 Volume 5 Issue 10, October 2017 www.ijser.in Licensed Under Creative Commons Attribution CC BY Improved Performance of Human Emotion Detection Using ECG Signal Processing Deepak Kumar 1 , Vivek Upadhyaya 2 1 M.Tech Scholar, Department of Digital Communication, Arya College of Engineering & IT, Jaipur, Rajasthan, India 2 Assistant Professor, Department of Electronics & Communication, Arya College of Engineering & IT, Jaipur, Rajasthan, India Abstract: Simulation and identification of emotions have attracted much interest from fields like cognitive science, psychology, and, recently, engineering. Even though a good quantity of investigation has been conducted on behavioral modalities, there are some under- researched aspects such as physiological signals. This research brings forth the ECG signal and introduces a complete study of its psychological characteristics. The very institution of this signal as a biometric property justifies subject-reliant emotion identifiers that record the immediate changeability of the signal from its homeostatic standard level. We are enhancing the implementation of the Emotion Identification through the application of ECG signals in this work. We recommend Ensemble Pragmatic Mode Decomposition or EPMD technique to diminish the operation duration and enhance the categorization rate. Keywords: Electrocardiogram, emotion recognition, affective computing, arousal, valence, active stress, passive stress, bivariate empirical mode decomposition, intrinsic mode function, instantaneous frequency, oscillation. 1. Introduction People’s emotions are psychophysiological occurrences that impact every facet of our day-to-day routine. Emotions are intricate operations composed of many elements, counting physical modifications, feelings, cognitive responses, actions, and reflections. Different models were recommended by taking into account the manners that these elements interact to generate emotions, though right now there is no unique method that is unanimously recognized. Simulating emotions is a really tough issue that has attracted a significant amount of attention from the developing area of human-computer interaction. The aim is to devise frameworks that may spontaneously recognize emotional phases, and which can modernize uses in the medical field, education, entertainment, protection, and so on. The major issue in devising such designs resides in our dependency on perceptible demonstrations of emotions to create and substantiate them as the dormant aspects that produce emotions are imperceptible. The primary phase in simulating any occurrence is information acquisition. We have to devise trials and introduce techniques that effectively incite emotions within a laboratory environment wherein we may save and acquire psychological information. When measuring psychological activity, we are restricted to the analysis of perceptible occurrences such as facial expressions, voice characteristics, gestures, and so on. Such modalities are common in HCI as they utilize similar prompts that people depend on to sense and identify emotional states. Furthermore, the majority of people show comparable manifestations as a consequence to similar emotional incitement, which permits for impartial emotion elucidation. One main disadvantage of utilizing behavioral modalities to sense emotions is the doubt that comes up for people who either are purposely controlling their emotional manifestations or are inherently emotionally suppressive. For example, even though facial expressions may be studied to deduce emotions, this does not assure that a person will manifest the equivalent prompting, regardless of whether they are feeling a particular emotion. This has grave impacts in some uses like observation. Physiological signals or biosignals are an attractive substitute to the utilization of behavioral modalities as they comprise of crucial signs of a person’s body. Instances in this classification count the electrocardiogram or ECG, electroencephalogram or EEG, electromyogram or EMG, galvanic skin response or GSR, heart rate (HR) or heart rate variability (HRV), blood volume pressure or BVP, respiration rate or RR, and temperature or T. Such signals have customarily been utilized for medical diagnostics, though there is consequent proof to imply that they are responsive to and can transmit data concerning emotional states [1-6]. Among the advantages of sensing emotions by utilizing biosignals is that they are the body’s reflexive responses, and hence are really tough to conceal. Furthermore, for the length of time that the detectors are connected to the body, such signals are saved constantly, permitting numerous emotional evaluations. This is contrary to voice characteristics, for instance, which may be recorded just when the person is talking. Though, there are numerous hypothetical and realistic issues with respect to physiological signal founded emotion sensing. Firstly, even though the proof implies that biosignals are impacted by emotions, the actual impacts on the waveform designs have yet to be observed. For instance, the heart rate rises both when someone is afraid and when excited, however, whether we may separate the two is till now obscure. Secondly, there are open queries concerning the subject-dependent aspect of these impacts. Aside from the open hypothetical queries, there are also realistic challenges. The preliminary conventions are much more intricate compared to behavioral emotion investigation, where the acquisition is helped by notifying subjects to show emotions. For physiological signals founded trials, more advanced procedures are required to Paper ID: IJSER171934 58 of 63