Take 2

An innovative Parkinson's diagnostic tool which combines typing analysis, voice patterns and blinking patterns to support early Parkinson's diagnosis, increasing accessibility to all.

Comment 4

There are ~90,000 new Parkinson's cases in the US every year, but diagnosing Parkinson's is a challenge because symptoms present subtly, and can often be missed in the initial stages. Our early Parkinson's diagnostic aid conducts keystroke, voice and eye blink analysis to be a crucial and early guide to detection of Parkinson's disease. We use VLMs for spectrograph analysis, LLMs for report generation on the patient's risk, where we give an overall risk score, and OpenCV for blink rate detection. We also took into mind ethical considerations regarding user data, agreeing to be HIPPA and SOC2 compliant, anonymise and not store user data. Our business strategy revolves around the target audience of people aged over 50, with our marketing strategy including TV advertisements, partnerships with clinics, old age homes and hospitals. We plan this to be a one time cost of $2 with a gross margin of 95% for us. Our costs for development and deployment would average ~$70 per month. The TAM for this market is ~1.9 billion people in the world aged over 50, while our SOM is 2 million people, initially, and we envision taking this to 200 million over time. Thank you!

Built With

Share this project:

Updates