Shyam Marjit, Upasana Talukdar, and Shyamanta M Hazarika.
Abstract: Emotion Recognition is an important problemwithin Affective Computing and Human Computer Interaction. In recent years, various machine learning models have provided significant progress in the field of emotion recognition. This paper proposes a framework for EEG-based emotion recognition using Multi Layer Perceptron (MLP). Power Spectral Density features were used for quantifying the emotions interms of valence-arousal scale and MLP is used for classification. Genetic algorithm is used to optimize the architecture of MLP. The proposed model identifies a. two classes of emotions viz. Low/High Valence with an average accuracy of 91.10% and Low/High Arousal with an average accuracy of 91.02%, b. four classes of emotions viz. High Valence-Low Arousal (HVLA), High Valence-High Arousal (HVHA), Low Valence-Low Arousal (LVLA) and Low Valence-High Arousal (LVHA) with 83.52% accuracy. The reported results are better compared to existing results in the literature.
Index Terms — EEG, Emotions, Power Spectral Density, Multi-Layer Perceptron, Genetic Algorithm.
✨ Description of few imp variables
dataset_path <- Path for the DEAP dataset
subject_no <- Subject number in string format for e.g. "s02", "s12"
generations <- No of generations in GA
pop_size <- Size of the population in GA
prob_cross <- Crossover probability in GA
prob_mut <- Mutation probability in GA
✨ Note for readers:
The new version of code provides better accuracy due to the updated code for preprocesing step, rest all are same as it was descrived in the paper. The performance of the new version of code are noted in the below table:
Valence | Arousal | 4-Types of emotions | |||||||
---|---|---|---|---|---|---|---|---|---|
Evolution Method | Accuracy | Precession | Recall | Accuracy | Precession | Recall | Accuracy | Precession | Recall |
Randomly splitted training and testing with 80:20 partitions |
92.2 ± 8.84 | 95.3 ± 12 | 84.5 ± 22.1 | 92.97 ± 8.95 | 93.18 ± 11.82 | 91.4 ± 17.1 | 85.94 ± 13.75 | 88.58 ± 12.43 | 85.94 ± 13.75 |
10-fold cross validation | 96.64 ± 2.43 | 96.02 ± 4.44 | 97.9 ± 4.7 | 96.56 ± 2.52 | 95.37 ± 4.82 | 97.2 ± 5.5 | 93.28 ± 2.87 | 90.76 ± 4.43 | 93.28 ± 2.87 |
If you think this project is helpful, please feel free to leave a star⭐️ and cite our paper:
@INPROCEEDINGS{shyam2021eeg,
title={EEG-Based Emotion Recognition Using Genetic Algorithm Optimized Multi-Layer Perceptron},
author={Marjit, Shyam and Talukdar, Upasana and Hazarika, Shyamanta M},
booktitle={2021 International Symposium of Asian Control Association on Intelligent Robotics and Industrial Automation (IRIA)},
year={2021},
pages={304-309},
organization={IEEE},
doi={10.1109/IRIA53009.2021.9588702}
}
Shyam Marjit: [email protected]