A feature extraction method based on ensemble median empirical mode decomposition achieved accuracy rates of 74.3% for valence and 78% for arousal classes in decoding EEG signals.
A feature extraction method based on MEEMD achieved accuracy rates of 74.3% for valence and 78% for arousal in emotion recognition using EEG signals.
This letter investigates ensemble median empirical mode decomposition (MEEMD), an extension model of ensemble empirical mode decomposition, and its improved characteristics for emotion recognition. It is tough to extract the hidden patterns in the electroencephalography (EEG) signal due to the signals' nonstationary nature, which is caused by the brain's complex neuronal activity. This makes it difficult to identify emotions using EEG. This research presents a feature extraction method based on MEEMD for decoding EEG signals for emotion recognition. Analysis is done on the intrinsic mode functions (IMFs) that are retrieved by EEMD and MEEMD. When identifying emotions using multichannel EEG signals, features like power spectral density, relative powers, power ratios, entropies, mean, standard deviation, and variance are used as indications of valence and arousal scales. The results indicate that the suggested method has achieved accuracy rates of 74.3% for valence and 78% for arousal classes. DEAP EEG emotion dataset is used, and both EEMD and MEEMD models are used to evaluate the results.
Samal et al. (Mon,) conducted a other in Emotion recognition. Ensemble median empirical mode decomposition (MEEMD) vs. Ensemble empirical mode decomposition (EEMD) was evaluated on Accuracy rates for valence and arousal classes. A feature extraction method based on ensemble median empirical mode decomposition achieved accuracy rates of 74.3% for valence and 78% for arousal classes in decoding EEG signals.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: