Respiratory Disorder Classification Based on Lung Auscultation sounds

Respiratory sounds are important indicators of respiratory health and respiratory disorders. The sound emitted when a person breathes is directly related to air movement, changes within lung tissue and the position of secretions within the lung. A wheezing sound, for example, is a common sign that a patient has an obstructive airway disease like asthma or chronic obstructive pulmonary disease (COPD). These sounds can be recorded using digital stethoscopes and other recording techniques. This digital data opens up the possibility of using machine learning to automatically diagnose respiratory disorders like asthma, pneumonia and bronchiolitis, to name a few.

Dataset Link

Respiratory Sound Database: https://www.kaggle.com/vbookshelf/respiratory-sound-database The dataset includes 920 annotated recordings of varying length - 10s to 90s. These recordings were taken from 126 patients. There are a total of 5.5 hours of recordings containing 6898 respiratory cycles - 1864 contain crackles, 886 contain wheezes and 506 contain both crackles and wheezes.

Docker Image for the app

• https://hub.docker.com/r/adityahb333/respiratory-disorder-classification
• Demo audio files for testing the application - https://github.com/adityahbapat/Respiratory-Disorder-Classification-Based-on-Lung-Auscultation-Sounds/tree/master/demo_audio_files

Requirements:

To get started with the project make sure you have python installed. (Backend folder has a different requirements.txt to support linux docker env) pip install -r requirements.txt

Folder Setup:

Here is a detailed folder setup to help you get started

respiratory_disorder_classification (Main Project Folder)

• Respiratory_Sound_Database (sub-folder)
• • audio_and_txt_files , etc
• venv (virtual env sub folder)
• backend (flask app)
• processed_audio_files
• .gitignore
• demographic_info.txt
• README.md
• respiratory_disorder_classification.ipynb
• training
• • train.csv
• validation
• • val.csv

creating a virtual env:

1. python -m venv C:\Users\"Aditya Bapat"\Desktop\project\respiratory_disorder_classification\venv (project folder path\venv)
2. give permission for windows user: Set-ExecutionPolicy -Scope CurrentUser -ExecutionPolicy Unrestricted
3. To activate: .\venv\Scripts\activate.ps1
4. Install the req packages: eg pip install tensorflow (without venv do: pip install tensorflow --user)

GPU support for tensorflow: https://www.youtube.com/watch?v=hHWkvEcDBO0

Flask App

Flask app can be found in the backend folder To run the app: python app.py