Usage

Setup

Please see the requirements.txt for environment configuration.

pip install -r requirements.txt

Processing

Please use the data_processing.py to process the MEG data.

cd ./src
python -u data_processing.py

Feature extraction

Please use the word_features.py and wav_features.py to extract the text and speech features respectively.

cd ./src
python -u word_features.py
python -u wav_features.py

Train and Test

Please use the following commands to train a model.

cd ./run_scripts
# Single-modal:
bash train_single.sh
bash train_cross_single.sh
## Our approach MASD: 
bash train.sh
bash train_phoneme.sh
bash train_cross.sh