ISSN: 2229-6948(ONLINE) ICTACT JOURNAL ON COMMUNICATION TECHNOLOGY, SEPTEMBER 2018, VOLUME: 09, ISSUE: 03 DOI: 10.21917/ijct.2018.0265 1815 EMOTION RECOGNITION BASED ON VARIOUS PHYSIOLOGICAL SIGNALS - A REVIEW Supriya Londhe 1 and Rushikesh Borse 2 1 Giesecke and Devrient MS India Private Limited, India 2 Electronics and Telecommunication Engineering, MIT Academy of Engineering, India Abstract Emotion recognition is one of the biggest challenges in human-human and human-computer interaction. There are various approaches to recognize emotions like facial expression, audio signals, body poses, and gestures etc. Physiological signals play vital role in emotion recognition as they are not controllable and are of immediate response type. In this paper, we discuss the research done on emotion recognition using skin conductance, skin temperature, electrocardiogram (ECG), electromyography (EMG), and electroencephalogram (EEG) signals. Altogether, the same methodology has been adopted for emotion recognition techniques based upon various physiological signals. After survey, it is understood that none of these methods are fully efficient standalone but the efficiency can be improved by using combination of physiological signals. The study of this paper provides an insight on the current state of research and challenges faced during emotion recognition using physiological signals, so that research can be advanced for better recognition. Keywords: Physiological Signals, Skin Conductance, EMG, ECG, EEG 1. INTRODUCTION Emotion is a psycho-physiological state of a human being which describes a person’s temperament. Although human emotional experience plays utmost important role in human lives, our scientific knowledge about human emotions has its own limitations. Hence, an ability to detect and recognize ones’ emotional state is essential in the improvement of artificial intelligence part of Human Machine Interaction (HMI). Emotion recognition is widely used in medical, defense, lie detection techniques, entertainment, education etc. Various findings from neuroscience, cognitive science and psychology suggest that emotions are important in human intelligence development, social interaction, perception, learning, and so on. Physiological signals are generated by the body during the functioning of various physiological systems. Hence, physiological signals hold information which can be extracted from these signals to find out the state of the functioning of these physiological systems. The process of extracting information can be very simple as feeling the pulse to find the state of heart beats and it can be complex which may require analysis of the structure of tissue by a sophisticated machine. Human beings understand each other’s emotion by means of text, speech and facial expressions, but there are many limitations to above said techniques, like emotion recognition using text is applied to words or sentences in a particular language so it is quite difficult to develop a universal system using text. Emotion recognition using facial expressions has many advantages but facial expressions are not always linked with inner emotions thus facial expressions can be consciously controlled. As a result, emotions cannot be recognized accurately. Emotion recognition using speech is much simpler than the above two but speech varies from person to person with different culture or geography hence speech recognition technique is also not feasible. To cope up with the limitations. Fig.1. Arousal/Valence Space Description of Emotions [16] For the above said methods, physiological signals can be used. Every human being generates various signals from different parts of the body. These signals are termed as physiological signals required for analysis of emotions using HCI [16]. There are many advantages of monitoring physiological patterns of the body controlled by the nervous system which is getting impacted by human emotions. There are various sensors which can be used for collecting physiological signals such as ECG, EMG, EEG, skin conductance, skin temperature and Blood Volume Pulse (BVP) [24]. The discrete emotional model claims the presence of some basic and universally standard emotions irrespective of culture or geography [16]. Several psychologists have suggested different categories of emotions. But there has been a considerable agreement in the following six emotions happiness, sadness, surprise, anger, disgust, and fear. The dimensional model categorizes emotions based on the scales and can be characterized on valance and arousal (Fig.1). Valance represents an emotional state, unpleasant to pleasant which varies from negative to positive. Arousal indicates the journey of human emotions calm to excite and travels from low to high. In this paper, we present a review of the recent advancements in emotion recognition techniques and corresponding research using physiological signals; in specific to the emotion elicitation stimuli, feature extraction and classification methodologies.