There is a wide range of feelings, thoughts, and behavior associated with emotions, as well as an outward manifestation of some of the more subtle aspects of human psychology. To feel an emotion is to experience the high-level psychological output of the human brain. It's an interdisciplinary field of study that draws from computer science, AI, neurology, healthcare, and other fields. People’s personalities and mental health are greatly influenced by their emotional experiences. In order to overcome stress and other stress-related diseases, one must be able to recognize and name one’s personal emotions. There has also been a lot of interest in the use of electroencephalogram (EEG) signals for emotion classification. In affective computing, EEG-based emotion recognition is a complicated and active research among researchers. The most important part in developing a highly effective brain-computer interface (BCI) system for emotion recognition using an electroencephalogram (EEG) are feature extractions and classifier selection. Recently, deep learning methods gained much importance for their better performance. In this paper, we propose a deep learning framework, Deep CNN for Emotion Recognition (DCNNER) that can detect human emotion with help of EEG signals. The use of principal component analysis (PCA) for feature selection and dimensionality reduction is also covered in this work. The suggested model uses principal component analysis (PCA) to minimize the dimensionality of the features and then feeds only the selected features to several classifiers. Different classifiers perform differently for pre-processed data and PCA applied data. Comparisons of different deep learning are discussed. The performance of the proposed system is measured with the model accuracy is 99% and model loss is 0.3.It is a 3-dimensional model which has valence, arousal and dominance for emotion detection.

EEG Signal, Stress, CNN, Deep Learning, PCA, Emotion Recognition, Accuracy