Welcome to the MMLung project, which is aimed at leveraging the power of smartphones for lung health estimation. Our primary script, encapsulated in the mmlung.ipynb Jupyter notebook, employs data from various sources to extract features and analyze different breathing tasks.
To successfully run the script, you'll need:
- Python 3.x
- pandas
- numpy
- glob
- ipywidgets
- IPython.display
- tqdm
Additionally, you'll need the pipe_scripts module in your Python path, which contains several functions used in our script.
Our mmlung.ipynb script is built to run in a Jupyter notebook environment and requires access to various data files. These include ground truth files and task files, which can be specified by setting the ground_truth_folder and tasks_folder variables respectively.
We also utilize a dictionary, cell_locations, to map the locations of various target variables within the data files. You may need to customize this dictionary to match the structure of your own data files.
The dataset for this project will be released in the future. In the meantime, you can record your own .wav files using any smartphone. Here are a couple of examples:
- Record a cough.
- Recite the Rainbow Passage: "When the sunlight strikes raindrops in the air, they act like a prism and form a rainbow. The rainbow is a division of white light into many beautiful colors..."
Ground truth data can be collected using the Easy on-PC PC spirometer. Find more information here.
MMLUNG System Pipeline
For a comprehensive understanding of our project, please refer to the mmlung.ipynb Jupyter notebook which contains detailed code and annotations.
Thank you for your interest in MMLung!
Copyright (c) 2022, University of Southampton All rights reserved. Check the License file for details.