Cancer is one of the highest forms of death among the global population and early detection of cancer dramatically increases chances for successful treatment. I utilized machine learning, artificial intelligence, and computer vision to help spot and stop cancer.

What it does

Skin Cancer Prediction - accuracy 84.7% Brain Tumor Prediction - accuracy 96.5%

How I built it

Combining a variety of Python libraries and Kaggle datasets(link on Github), each model was trained on data, and the regression equations were exported then the equation was used to make a prediction about a user-defined piece of data. All datasets were preprocessed and split into test and training sets, and each model's accuracy was verified against the test sets.

Challenges I ran into

Preprocessing data, imputing missing values, standardizing data, installing the necessary libraries, managing run-time issues, transferring large datasets, overfitting, and linking the back end with the front end.

Accomplishments that I'm proud of

I am proud of the fact that all of my predictions were made with models that I have trained by ourselves, and no pre-trained ML software was used at all. In addition, I've been able to utilize our different educational backgrounds to bring unique insights to the project. Most importantly however, is that this product serves to combat cancer and has the potential to possibly save a life.

What I learned

Using Cloud services to train the model and working with Keras library We learned how to use machine learning algorithms for the back end and later applying them on apps

What's next for CancerPrediction

A next step that could be taken with this project is potentially pairing it with a mobile app so that photos can be taken by users and predictions can be made in real-time about the photos in question. The back end code is pretty much complete, so if the regression equation is exported and paired with the front end code, a skin cancer detection app could be very possible.

Share this project: