Describe about the Emotion Recognition by Physiological Signals Brain and EEG Signals?
Recent research on Human Computer interaction focuses at the recognition of user’s emotional state to provide a reliable interface between computer and humans. This concept would provide easy life o deal with. There is a vast application involving such areas of medicine, education etc. Human emotion is recognizable by verities of approaches such as facial images, gesture, neuro imaging methods and physiological signals. Various theories on emotion recognition and current advancement of their methodologies have developed and still going on.
There are so many ways to record psychophysiology data from humans, e.g. Electromyography (EMG), Electromyography (EMG) and Electrocardiogram (ECG). Here being EEG as the focus point various techniques on machine learning has evolved, such as, K-Nearest Neighbor (KNN), Artificial Neural Network (ANN), Bayesian Network (BN) and Support Vector Machine (SVM), these are some techniques on machine learning to classify the EEG data being used in varieties of experiments (Murugappan, Murugappan and Zheng, 2013).
Application of EEG:
Most of the developed methods in the literature of neuro physiological studies have submitted the report o the correlation between emotion states and EEG signals. All these methods were based on frequency-domain analysis and time-domain analysis. Event related potentials (ERPs) components reflect the emotional states. To have correlation with the valence, the ERP components of having short to middle latencies have been shown, whereas to have the correlation with the arousal there are the ERP components having middle to long latencies. With the computation of the ERP signals, it requires EEG signal averaging over multiple trials. It is inappropriate to render ERP feature for the online processing (Panayiotopoulos, 2005).
Classification of emotions is not deterministic, they are probabilistic in nature. Previous researches over the human emotion recognition has dealt with the probability theory driven classification for the estimation of Human bandwidth of 0.5-100 Hz has been in use to remove noise. Here application of 50 Hz notch filter is done in order to remove noise for power-line interface. Preprocessing including generated vectored input stream, pre-filtering is required in Multi Wave Transform (MWT). There are different possible ways are available to gain the vector input stream. In this task, Vector Input Stream was obtained with the use of repeated scheme of row pre-processing. The filter bank with matrix-valued wavelet also requires inputs with multiple stream, decided by the multiplicity (Petrantonakis and Hadjileontiadis, 2010).
All the previous attempts on emotion recognition with the help of the EEG signals are primarily focused on collection of data from nervous system. After going through some literature survey, this brought out two divided opinions. One group of researchers gained less accurate emotion classification using EEG signals, where they state that physiological signals like heart rate, galvanic skin resistance, respiration etc. were in use to get high accuracy. On the other hand they stated their view that EEG signal is sufficient alone for emotion classification, though they are facing it less accurate. It is proposed that only EEG signals alone is used to classify the emotions using neural network two state classifiers, where the selection of features are made for the improvement of the accurate classification (Selvaraj et al., 2013).
Murugappan, M., Murugappan, S., & Zheng, B. (2013). Frequency Band Analysis of Electrocardiogram (ECG) Signals for Human Emotional State Classification Using Discrete Wavelet Transform (DWT). Journal Of Physical Therapy Science, 25(7), 753-759. doi:10.1589/jpts.25.753
Panayiotopoulos, C. (2005). Optimal Use of the EEG in the Diagnosis and Management of Epilepsies. Bladon Medical Publishing. Retrieved from https://www.ncbi.nlm.nih.gov/books/NBK2601/
Petrantonakis, P., & Hadjileontiadis, L. (2010). Emotion Recognition From EEG Using Higher Order Crossings. IEEE Transactions On Information Technology In Biomedicine, 14(2), 186-197. doi:10.1109/titb.2009.2034649
Selvaraj, J., Murugappan, M., Wan, K., & Yaacob, S. (2013). Classification of emotional states from electrocardiogram signals: a non-linear approach based on hurst. Biomed Eng Online, 12(1), 44. doi:10.1186/1475-925x-12-44