- This is the repo to turn OpenBCI raw data into spectrograms and to train the model on said data.
- Checks the latest recordings in the SavedData folder from OpenBCI, takes the array of values of each channel
- and turns it into a wavfile.
- Before executing this code you first need to split the data using the
process_audio.py
fromsubvocalization_stream
, - once you split the wavfiles, this code transforms the individual wavfiles into sepctrograms.
- This script will get the file location of each spectrogram and list them into a csv file for training the model.
- This script will train the model using the assigned csv file.
- Using the trained model we can apply it to the dataset to find outliers and create a csv file to exclude said outlier data.