Emotion-Recognition-Using-ECG-Data

Download

• Dataset
• Preprocessed data

Requirements

• Python =3.7.9
• TensorFlow = 1.15.0
• TensorBoard = 1.15.0
• NumPy = 1.19.2
• Pandas = 1.2.2
• Scikit-Learn = 0.24.1
• Tqdm = 4.50.2
• Mlxtend = 0.18.0

Code structure

• logs/logs_kfold : store the log files during training.
• output/output_kfold : store the performence evaluation files (tr means train performence, te:test)
• weights/weights_kfold : store the weight of trained wesad_model# in wesad_model#.h5 format.
• PreprocessedData : store the preprocessed data obtained by runing data_preprocess.py; if you want to run data_preprocess.py, modify "preprocess_data = 0" to "preprocess_data = 1".
• model.py : build model structure; we have 5 models here.
• kfold.py : implement kfold cross-validation on models; compare and choose the best model.
• DL.py : train the best model.

Implementation

• Change the "model_name" in kfold.py/DL.py; you could choose 'wesad_affect1'/'wesad_affect2'/'wesad_affect4'/'wesad_affect5'/'wesad_affect6'.
• run DL.py