Cough Classification Model

Overview

This repository contains the code for a machine learning model developed to classify cough sounds by differentiating between wet and dry sounds. The project utilizes the Librosa and TensorFlow libraries for audio processing and machine learning, respectively. The data, sourced from Kaggle, is visualized using IPython to generate spectrograms, providing insights into the characteristics of the cough sounds.

Features

• Audio Classification: The model is designed to distinguish between wet and dry cough sounds.
• Librosa and TensorFlow: Librosa is employed for audio feature extraction, while TensorFlow is used for building and training the machine learning model.
• IPython Visualization: Spectrograms are created using IPython to visually represent the audio data, aiding in the understanding of the cough sound characteristics.

Results

The classification model achieved a 70% accuracy rate in differentiating between wet and dry cough sounds. The results are based on the evaluation of the model using a specified dataset.

Contribution to Society

The project's primary contribution lies in its potential applications for public health and epidemiological monitoring. By accurately classifying cough sounds, the model could be integrated into healthcare systems for early detection of respiratory conditions, such as identifying symptoms associated with infectious diseases. This early detection capability has the following societal benefits:

• Early Disease Identification: The model can contribute to the early identification of respiratory conditions, allowing for timely intervention and treatment.
• Epidemiological Monitoring: Monitoring and analyzing cough patterns in different regions can provide valuable insights for public health officials in tracking the spread of infectious diseases.
• Resource Allocation: Improved disease surveillance can aid in more efficient allocation of healthcare resources during outbreaks, ensuring that affected areas receive timely support.

Usage

1. Clone the repository:

   git clone https://github.com/brilboy/cough-classification.git
   cd cough-classification

2. Install the required dependencies:

   pip install -r requirements.txt

3. Open and run the Jupyter notebook:

   jupyter notebook cough_classification.ipynb

4. Follow the instructions within the notebook to train the model, visualize the data, and evaluate the results.

Data Source

The dataset used in this project is sourced from Kaggle. Special thanks to the contributors for providing this valuable resource.

Dependencies

• Librosa
• TensorFlow
• IPython

Acknowledgments

• The project makes use of the powerful Librosa and TensorFlow libraries for audio processing and machine learning.
• IPython is utilized for interactive and informative data visualization.

Feel free to explore the notebook and adapt the code for your own audio classification projects. If you encounter any issues or have suggestions for improvement, please open an issue or submit a pull request. Happy coding!