This repository contains the tensorflow implementation for the paper: "Emotion Recognition from Multi-Channel EEG through Parallel Convolutional Recurrent Neural Network"
- Title: Emotion Recognition from Multi-Channel EEG through Parallel Convolutional Recurrent Neural Network
- Authors: Yilong Yang, Qingfeng Wu, Ming Qiu, Yingdong Wang, Xiaowei Chen
- Institution: Xiamen University
- Published in: 2018 International Joint Conference on Neural Networks (IJCNN)
- DOI: 10.1109/IJCNN.2018.8489331
- Before running the code, please download the DEAP dataset, unzip it and place it into the right directory. The dataset can be found here. Each .mat data file contains the EEG signals and consponding labels of a subject. There are 2 arrays in the file: data and labels. The shape of data is (40, 40, 8064). The shape of label is (40,4). Each .pkl file contains a numpy.ndarray variable. It stores the pre_processed data with the shape of (segments, window_size, width, height), in this paper, it is (2400,128,9,9).
- Please run the deap_pre_process.py to Load the origin .mat data file and transform it into .pkl file.
- Using cv.py to train and test the model (10-fold cross-validation), result of each fold will be saved in a .xls file.
- count_accuracy.py is used to caculate the final accuracy of the model.
- Pyhton 3
- scipy
- numpy
- pandas
- pickle
- sk-learn
- pickle
- tensorflow (1.4 or higher version)
- xlrd
- xlwt