Parkinson's Machine Learning Predictor

A program that uses a user-drawn image of a spiral to predict if they have Parkinson's disease through the use of a trained dataset with high accuracy!

Comment

Our inspiration for this project was to combine the knowledge we acquired through neuroscience and link it to a project focused on computer science. Our goal is to make life easier for individuals questioning if they might have Parkinson's disease. The model that we built helps someone objectively test their hand movements in comparison to healthy controls. Even though this project is in its early stages, we believe that it can help many individuals out, especially the older population.

The older populous is rarely the focus of disease research and more attention needs to be driven to these patients. If something is accessible at home, an individual of any age will want to use this tool to potentially find out any risks they might have. We want to clarify that this model will "predict" and not diagnose a patient with the disease. It would be a good stepping stone in encouraging the patient to visit their provider. Also, it is a fun way to interact with technology and understand how powerful machine learning is.

Our model is able to take in an image of a spiral drawn by the user and show a resulting prediction based on the trained dataset.

We built it using several packages in Python such as openCV, tinker, numpy, and much more. We built it by making a backend and then developing a frontend for user accessibility.

In the beginning, we were struggling to understand if the dataset is predetermined and all the future predictions are based on that. Then, we decided that we would like the model to be real-time and build upon itself every time the user uses the application. Some of the packages were also new to us such as openCV and tinker, but we researched the functions online to help us achieve our goals.

We are proud of being able to make something useful in the health sector. We are proud of being able to combine our knowledge of neuroscience and make an idea into reality by executing a very technical project.

We learned machine learning as a whole considering neither of us was an expert when starting this. Actually, we are beginners considering this is our first hackathon. We learned a lot more about how to use GitHub, python packages, and how to make an interactive user interface.

In the future, we hope to make more changes and additions to our project for the betterment of this disease. We will work with the pre-health department at GT to bring this to the masses and be used as a regular tool in the healthcare setting.

To learn more about Parkinson's disease, please check out this link: https://www.nia.nih.gov/health/parkinsons-disease

To learn more about this project's background information, please check out this link: https://youtu.be/lGLIArxX6IA

Built With

Share this project:

Updates