Abstract:The recognition of human emotions based on physiological signals has garnered significant attention from various fields such as human–computer interaction, cognitive-behavioral science, and the treatment of emotion-related diseases. Using the easily accessible ECG signal for emotion recognition presents a viable approach. However, method relying on hand-crafted features necessitate task-specific domain knowledge and consume substantial time to design suitable features, and studies employing deep learning technologies incur greater computational complexity and extended training durations. To address these problems, we proposed an emotion recognition method that employs random convolutional kernels on ECG signals. Following ECG signal preprocessing, a large number of random convolutional kernels were used to transform the signals, and the resulting features are utilized to classify the various emotion states. In addition to accurately and automatically extracting the multi-scale features from ECG signals, combining the random convolutional kernels with ECG signal alone also reduces the computational complexity and training time compared to methods using multiple physiological signals or deep neural networks. Validation on three publicly available datasets revealed that the proposed method achieved recognition accuracies of 91.5%, 96% and 92.5% for valence, arousal and dominance in AMIGOS dataset, and 95.1%, 98.8% and 90.2% for same labels in DREAMER dataset. The accuracy for valence, arousal and affective states in WESAD dataset are 94.5%, 91.8% and 91.8%. The results show that the proposed method outperforms previous researches in multi-class emotion recognition, thus offering a new and promising approach for swift and precise emotion recognition.

ECG-based emotion recognition using random convolutional kernel method

A Efficient Multimodal Framework for Large Scale Emotion Recognition by Fusing Music and Electrodermal Activity Signals

User Independent Emotion Recognition with Residual Signal-Image Network

Using electroencephalogram for emotion recognition based on filter-bank long short-term memory networks

A new deep convolutional neural network incorporating attentional mechanisms for ECG emotion recognition

Emotion recognition with convolutional neural network and EEG-based EFDMs

EEG-Based Emotion Recognition by Convolutional Neural Network with Multi-Scale Kernels

EEG-Based Emotion Recognition Using a 2D CNN with Different Kernels

Electroencephalogram Emotion Recognition Based on Empirical Mode Decomposition and Optimal Feature Selection.

Emotion Recognition using Machine Learning and ECG signals

A novel one-dimensional convolutional neural network-based method for emotion recognition of electric power industry workers

A 3D Convolutional Neural Network for Emotion Recognition Based on EEG Signals.

Emotion Recognition Using Electrodermal Activity Signals and Multiscale Deep Convolution Neural Network

Research on Emotion Recognition of EEG Signal Based on Convolutional Neural Networks and High-Order Cross-Analysis

A new data augmentation convolutional neural network for human emotion recognition based on ECG signals

EEG-based emotion recognition using an end-to-end regional-asymmetric convolutional neural network

EEG emotion recognition approach using multi-scale convolution and feature fusion

ER-GET: Emotion Recognition Based on Global ECG Trajectory

Electroencephalograph-based emotion recognition using convolutional neural network without manual feature extraction

Deep Convolutional Neural Network For Emotion Recognition Using Eeg And Peripheral Physiological Signal

Research on emotion recognition based on ECG signal