Homayoon Yektaei - Master of biomedical engineering, Department of Biomedical Engineering, Islamic Azad University, Tehran North Branch/Tehran,Iran
Hanieh Yektai - biomedical engineer, Department of Biomedical Engineering, Ahrar University

Introduction: in this research, we have shown emotion recognition through EEG processing. Inthe beginning, the general definitions of the term are to further study the structure of the humanbrain as the brain signal generator, and then we will explain the electroencephalogram.Methods: In this study, some of the most important features of the extraction feature arementioned. These described is DWT- PCA- DFT- STFT- EMD methods include linear andnonlinear methods or analyzes in time domain and frequency. One of the linear methods we haveAnd non-linear methods can be pointed out RP- PP- ICA. Finally, the accuracy and precision ofthe operation of each of the most important categories are stated for the classification of the generaland final categorization. In this study, we describe the classification methods SVM- KNN - NN -LDA- QDA case We reviewed.Results: Neural networks also had easy training and careful classification. The accuracy of theclassification function performance was reported using the 48.78% neural network. k- nearestneighbor was easy to understand and easy to implement, but it worked poorly at runtime. Theaccuracy of this type of classification is 52.44%. In the research the results of classification withbackup vector machine 56.10% reported.

Emotion Recognition, EEG signals, Feature Extraction, Linear methods, non-linear methods, Classification