It is very much evident what Deep Learning can achieve in the Medical field, hence in view of the current situation of pandemic and for education purpose, I have created a Web Application that could detect Covid-19 in X-ray images. Obtained following performance metrics:
- Accuracy: 98%
- Sensitivity: 95%
- Specificity: 100%
pip install the following libraries:
- pandas
- numpy
- scikit-learn
- tensorflow
- keras
- flask
- opencv-python
- Pillow
Dataset: 1) link for Covid-19 X-ray images 2) link for Healthy persons X-ray images
Above datasets can be found in this repository in the Covid-19_model/covid-chestxray-dataset-master folder (For Covid-19 patients) and in the Covid-19_model/chest_xray folder (For Healthy persons).
covid_positive_dataset.py and kaggle_normal.py files are used to extract images from the above mentioned datasets folder and copy it in Covid-19_model/dataset folder.
covid_model.py file contains the actual code of model creation and covid_initial.h5 contain the saved model.
For running the Web application, run the predict_app.py (Flask application) file in the flask_app folder.
This is how the Web Application looks and work.
