Inspiration
The journey of AlziAid began with a compelling need to address the limitations of traditional questionnaire-based Alzheimer's disease (AD) diagnosis. Current methods often lack the precision and early detection capabilities necessary for effective treatment planning. This gap in healthcare technology inspired the development of AlziAid, a program designed to transform AD diagnostics through face landmarking and image recognition technologies.
What it does
AlziAid utilizes real-time face landmark extraction and surface interpolation to analyze iris movements for AD diagnosis. By leveraging Google's MediaPipe and Accelerate framework, it identifies subtle differences in iris movement between healthy individuals and AD patients, which are not detectable in conventional diagnostic methods (e.g., questionnaires).
How we built it
Our approach integrated Google's MediaPipe for facial landmark extraction and used surface interpolation techniques for mapping facial structures. We employed the Accelerate framework for detailed analysis and built AI models focusing on iris movements, hypothesizing slower vertical movements in AD patients compared to healthy individuals.
Challenges we ran into
Ensuring the accuracy and reliability of our AI models was a significant challenge, given the subtle differences in iris movements. Additionally, ethical considerations and privacy concerns with facial recognition technology were critical issues we addressed with a focus on anonymized data and privacy compliance.
Accomplishments that we're proud of
Developing an innovative, AI-assisted diagnostic tool that has the potential to revolutionize Alzheimer's disease diagnosis is our proudest achievement. Our integration of advanced technologies in a user-friendly application represents a significant step forward in healthcare diagnostics.
What we learned
We gained insights into the complexities of facial recognition technology, the ethical and privacy aspects of using such technology in healthcare, and the intricacies of building accurate and reliable AI models for medical diagnosis.
What's next for AlziAid
We plan extensive validation tests with 100 subjects divided equally between AD patients and healthy controls. We aim to improve AlziAid's accuracy and integrate it with traditional diagnostic methods, contributing to better early detection and treatment strategies for Alzheimer's disease.
Log in or sign up for Devpost to join the conversation.