WomenWhoCode Hackathon Covid Detection and Antibiotic Resistance Prevention application.
The application can be found here.
To run the application clone the repository and follow these steps :
- Clone the repository to your local machine.
- Create a new python virtual environment.
- Got to the Deployment directory and open terminal.
- Run pip install -r requiremnts.txt to install the required dependencies.
- Run streamlit run app.py to run the application.
[IEEE8023 GitHub Account] [Mendeley]
- DenseNet121. The pre-trained files were obtained from the course 'AI for Medical Diagnosis', in Coursera
- AWS
- AWS Sagemaker
- AWS S3
- Streamlit
- Heroku
The project demonstrates the Covid Detection and Antibiotic Resistance Prevention application built into an interactive Streamlit app.
- Random Selection from ZhangLabData.ipynb
Function: From thousands of ‘Normal’ and ‘Bacterial Pneumonia’ images in Mendeley, this code randomly selects 490 for each class for our project. - Dataset_Creation_Final.ipynb
Function: It splits all images into training, test and validation sets using stratified random sampling in order to ensure that all sets come from a similar distribution.
- SageMaker_Model.ipynb
Function: It generates our model in AWS SageMaker. - densenet.hdf5
Function: These are the pre-trained weights, obtained from the course 'AI for Medical Diagnosis', in Coursera.
- model_architecture.json and model_weights.h5
Function: These are our trained model files. - model_prediction.py
Function: It generates predictions using our model for the X-rays uploaded in the streamlit application. - app.py
Function: This is the main streamlit application that generates prediction for an uploaded X-ray image and also provide antibiotic resistance guidelines. - requirements.txt
Function: It contains all the package versions used for deployment.