Emotion Recognition Using Electroencephalogram Signals and a 1D Local Binary Pattern for an ML-Classification-Based Approach

Machine learning techniques provide low complexity in classification algorithms and feature analysis. In the emotion recognition area, these techniques have reduced their impact with the development of deep learning. Taking advantage of low-cost machine learning models, in this study, we introduce a novel method for emotion recognition using electroencephalographic signals and a local binary pattern to extract the periodicity of a signal and apply a feature extraction algorithm. Five emotions are classified with a support vector machine using a feature matrix reduced using a minimum redundancy maximum relevance algorithm, obtaining the first, homogeneity, and spectral roll-off as the best emotion descriptors. The resulting model is evaluated using accuracy, recall, precision, and F1-score for 10-cross-validation folders. Achieved results are compared with the existing literature showing high performance in the machine learning approaches.