Our journey began with a deep compassion for those affected by Parkinson's disease and a genuine desire to make a difference in their lives. Witnessing the struggles and challenges faced by individuals living with this neurological disorder, we were inspired to explore innovative ways to support them on their path to wellness.

The motivation to develop ML models for Parkinson's detection stemmed from a shared vision of early intervention and improved outcomes. We believe that every individual deserves the best possible chance at managing their health, and by leveraging the power of machine learning, we saw an opportunity to provide early detection tools that could potentially transform lives.

The immense potential of artificial intelligence and ML algorithms to analyze complex data and identify subtle patterns further fueled our inspiration. We envisioned a future where cutting-edge technology could assist healthcare professionals in identifying the earliest signs of Parkinson's, allowing for proactive treatment plans that mitigate the progression of the disease and alleviate symptoms.

Moreover, our project's inclusion of an AI virtual whiteboard was driven by a genuine desire to bridge the gap between patients and healthcare providers. We sought to create a collaborative and intuitive platform where individuals could give tests without the hassle of going through multiple layers to upload their projects. By fostering a supportive and interactive environment, we aimed to enhance the overall well-being and quality of care for Parkinson's patients.

In essence, our inspiration stems from empathy, a yearning to make a positive impact, and a deep-rooted belief in the transformative power of technology. We are driven by the profound motivation to improve the lives of those affected by Parkinson's disease, providing them with the tools and support they need to navigate their journey with hope, dignity, and resilience.

What it does

Parkinson's Risk Evaluator is more than just an online tool; it is a compassionate and user-friendly platform designed to empower individuals in assessing their risk of Parkinson's disease. We understand that timely detection is crucial for effective management of the condition, and our inspiration comes from the desire to provide a convenient and accessible solution for early risk assessment.

Through the Spiral Drawing test, users can simply submit pictures of spirals they draw, which are then meticulously analyzed by our advanced ML models. These models are trained to identify patterns present in the drawings that are indicative of Parkinson's disease. Similarly, the Voice-based test captures a brief 6-second voice sample from the user, using AI algorithms to detect subtle markers in speech that correlate with the condition.

It is important to emphasize that while these tests provide valuable insights, they do not replace the need for an official diagnosis from medical professionals. We recognize the vital role of doctors in accurately diagnosing Parkinson's disease. To ensure our users are well-informed, the app not only offers test results but also provides comprehensive information about Parkinson's disease. Users can gain a deeper understanding of the condition, its symptoms, and available treatment options.

In our commitment to support individuals, we offer detailed explanations about the tests themselves, clarifying the purpose and methodology behind each assessment. We believe in empowering users through knowledge, so they can make informed decisions and seek appropriate medical guidance.

For those who receive concerning test results, the app goes a step further by providing additional resources. We understand the emotional impact of potentially alarming results and aim to connect users with support networks, patient communities, and organizations dedicated to Parkinson's disease. These resources ensure that individuals can access further assistance and find solace in knowing they are not alone in their journey.

At its core, Parkinson's Risk Evaluator is a humane and caring initiative. We strive to make a positive impact on the lives of individuals by offering a user-friendly platform for risk assessment, accurate information about the disease, and a compassionate support system. By combining technology, empathy, and comprehensive resources, we hope to create a space that promotes early detection, understanding, and empowerment for those affected by Parkinson's disease.

In addition to the Parkinson's Risk Evaluator, we are proud to introduce our AI Virtual Whiteboard feature, which aims to enhance collaboration and communication among individuals and healthcare professionals in the context of Parkinson's disease.

How we built it

We created the web app with a HTML/CSS/Javascript frontend and Python/Flask server-side, which was used to run the image and voice processing machine learning algorithms written with TensorFlow on Python.

Challenges we ran into

To bring the Parkinson's Risk Evaluator and AI Virtual Whiteboard to life, we utilized a combination of various technologies and frameworks, including HTML, CSS, JavaScript, TensorFlow, Flask, and OpenCV.

The front-end of the application was built using HTML, CSS, and JavaScript. HTML provided the structure and layout of the web pages, CSS was used to style and enhance the visual presentation, and JavaScript added interactivity and dynamic functionality. This combination allowed us to create an intuitive and user-friendly interface for users to interact with the application seamlessly.

For the ML models used in the Spiral Drawing and Voice-based tests, we leveraged TensorFlow, a powerful open-source machine learning framework. TensorFlow provided the foundation for training and deploying the ML models that analyze user input. Through extensive training on datasets containing examples of Parkinson's disease markers, the models can detect patterns and make predictions based on new inputs.

To integrate the application's front-end with the back-end, we employed Flask, a Python-based web framework. Flask allowed us to develop the server-side logic and APIs needed for communication between the user interface and the ML models. It provided a flexible and scalable framework for handling user requests, processing data, and generating responses.

In the case of the AI Virtual Whiteboard, we utilized OpenCV, an open-source computer vision library. OpenCV enabled us to incorporate AI-driven features into the whiteboard, such as handwriting recognition and image processing. These features enhance the functionality of the whiteboard by allowing users to transcribe and manipulate handwritten text, as well as analyze and annotate images.

By combining HTML, CSS, JavaScript, TensorFlow, Flask, and OpenCV, we were able to create an integrated and robust application. The seamless integration of these technologies enables users to conveniently assess their risk of Parkinson's disease, access relevant information, and collaborate effectively through the AI Virtual Whiteboard.

Accomplishments that we're proud of

  • Georgiy implemented a logistic regression model based on .wav files which proved to be more difficult than initially anticipated. The most complicated part was extracting about 10 different characteristic information from .wav files.
  • Melody designed the website without any external libraries from scratch. She made the logo from scratch which is a combination of flower (representative of the disease) along with a spiral base (representing the test).
  • Shyam made a point about having great code quality this time compared to other projects.
  • Sajid integrated Georgiy's models to the front end using Flask which was pretty challenging.

What we learned

We learned that for the most efficient use of time, development environments should be set up and verified much earlier.

Built With

Share this project: