NOTE; The project is aimed at using histological features, CNN and classifiers, currently anatomical features is been used, This page will be updated as soon as the project is completed.

Breast cancer is the second most common cancer among women in the United States and the leading cancer type in Nigeria Source.

However, it can be prevented or managed by early diagnosis. Radiography and Palpation (observation) are the common method in 3rd world countries like Nigeria and other African countires.

Furthermore, insufficient medical personnels (due to mass exodus of Nigerian doctors to the U.K) & lack of efficient medical facilities, early diagnosis is rare.

Therefore, A.I as a leading tech can be used to automate diagnosis process by:

• Getting a large and robust dataset (check kaggle for anatomical model dataset)
• Training model with the data (I used random forest for anatomical features)
• Deploying, Evaluating and tuning Model. (I used streamlit, github and Heroku).

Four parameters are put into considerations as parameters due to there correlation with the target features. The features are:

• Radius_mean
• Perimeter_mean
• Area_mean
• Concave points_mean.

Toggle with parameters and the model will diagnose whether the cancer is Benign or Malignant 0-0.5 is Benign 0.55 - 1 is Malignant.

Again, the histological features will be added as soon as the CNN is completed.

Click here to check the Web App

Built and deployed with Pickle, Streamlit and Heroku

THANK YOU