1. Automated accurate emotion classification using Clefia pattern-based features with EEG signals.
- Author
-
Dogan, Abdullah, Barua, Prabal Datta, Baygin, Mehmet, Tuncer, Turker, Dogan, Sengul, Yaman, Orhan, Dogru, Ali Hikmet, and Acharya, Rajendra U.
- Subjects
SIGNAL classification ,ELECTROENCEPHALOGRAPHY ,PLURALITY voting ,EMOTIONS ,AUTOMATIC classification - Abstract
This article discusses the challenges of manually screening electroencephalogram (EEG) signals for emotion classification and proposes a novel automated method using Clefia pattern-based features. The proposed model achieves high accuracy in classifying arousal, dominance, and valence cases using two public databases. The Clefia pattern-based method offers low computational complexity and high accuracy for automated emotion classification using EEG signals. The text describes a database called DREAMER that contains EEG and ECG signals collected from 23 participants. The text also explains a proposed model for emotion classification using Clefia pattern and mRMR feature selection. The given text describes a method called Clefia pattern for extracting features from EEG signals. The method involves dividing the EEG signal into overlapping blocks and comparing each value in the signal with a corresponding value in a binary array. The resulting binary array is then divided into two groups and the decimal equivalent of each group is calculated. Finally, histograms of the EEG signal are generated using these decimal values and a feature vector is obtained by combining the histograms. The Clefia pattern is shown in a pseudocode algorithm and sample results of map generation and histogram operations are provided. The text describes a model for emotion classification that involves feature selection, classification using a support vector machine (SVM), and majority voting. The model is validated using two public databases and achieves high classification performance for arousal, valence, and dominance emotional states. The results are presented in tables, showing the performance metrics for each channel [Extracted from the article]
- Published
- 2024
- Full Text
- View/download PDF