This project a MFCC-based Recurrent Neural Network for automatic clinical depression recognition and assessment from speech model deployed as a Web Application. The model is a based on one of the proposed solutions of the paper Rejaibi, Emna, Komaty, Ali, Meriaudeau, Fabrice, & Othmani, Morgan-Hiring. (2022). MFCC-based Recurrent Neural Network for Automatic Clinical Depression Recognition and Assessment from Speech. Biomedical Signal Processing and Control, 71, 103107. doi: 10.1016/j.bspc.2021.103107.
The pre processing function and the model have been is slightly altered . The Dataset used is the DIAC-WOZ dataset
- Python (^3.10) [with PIP]
- NPM (^9.50)
- Node.js (^16.16.0)
- React (^18.2)
- Git
Clone the repository by downloading the zip file or running the git clone command like so:
git clone https://github.com/RamgopalH/DepressionDetectionUsingLSTM.git
Change to the server folder and do the following:
To set up the python packages, run the followign command o install all pytohn depemdencies
pip install numpy pandas pydub IPython librosa scipy tensorflow
Install All the required node modules using npm
npm init -y
npm install
Make sure port 5000 is free on your system or change the value of PORT in the index.js file
To start the server, run
nodemon index.js
Switch to the client folder and set up the frontend of the application
Install all required Node packages and React packages using NPM
npm init -y
npm install
Make sure sever is running before running the client with the following command
npm run dev
It displays a url in the command prompt and following this url should take you to the home page of the application.