Table 1 State-of-the-art machine learning techniques for memory-induced emotion recognition using EEG signals with dataset information, compared with proposed technique and dataset.
From: Insights from EEG analysis of evoked memory recalls using deep learning for emotion charting
Study | Method | Evoked memory technique | Modality | Classes | Subjects |
|---|---|---|---|---|---|
Chanel et al.22 | Temporal and frequency domain features with Linear discriminant analysis (LDA) classifier | Memory recall relevant to personalized stimulus images | EEG (62 channels) | Three classes (positive, negative, neutral) | 10 |
Iacoviello et al.23 | Wavelet transform feature extraction, Principal component analysis for feature selection, and SVM for classification | Memory recall of unpleasant odors | EEG (8 channels) | Two classes (Disgust or not disgust) | 10 |
Zhuang et al.24 | Differential entropy features, and SVM for classification | Memory recall of recently displayed video stimulus | EEG (62 channels) | Six basic emotions | 30 |
Proposed | One dimensional convolutional recurrent neural network with combination of extreme learning machine (1D-CRNN-ELM) | Memory recall with displayed words | EEG (14 channels) | Four emotion classes (HVHA, HVLA, LVHA, LVLA) | 69 |
