My teammate's (Dylan) grandmother, with whom he was very close, was diagnosed with Alzheimer's disease when he was just 13 years old. It was extraordinarily crushing for Dylan since his grandmother was very soon unable to recognize him. Seeing his grandmother in that kind of a state forged in Dylan a strong will to fight Alzheimer's disease and improve the lives of those affected by it and their loved ones. We hope that our product will be able to help those affected by Alzheimer's to be able to recognize and remember their loved ones.
What it does
Mindeye first allows the user to take a picture of a loved one or friend. After uploading the picture to our website, the user inputs the relation they have with the individual whose picture was taken and their name. After this is completed, the user will be able to see that their name and relation to the Alzheimer's patient show up in a live webcam run.
How we built it
We built mindeye as an OpenCV Python app enhanced with friendly UI through a frontend web server. Through OpenCV's heatmap detection capabilities, we were able to identify faces in a picture. Furthermore, using facial recognition technology, we were able to process key features in facial structure to quickly identify users in live video. Finally, our web server contains a live webcam and uses PHP to allow users to insert their own facial profile and relation to the Alzheimer's patient conveniently and without having to access any of the backend.
Challenges we ran into
- Learning OpenCV to annotate live video in one day
- Incorrect facial identification in the beginning (intense 3 hour debugging session)
- Styling (God damn the styling...)
Accomplishments that we're proud of
- Learning OpenCV to annotate live video
- Running a VERY effective machine learning facial recognition
- Creating a beautiful GUI on the user end to allow users to quickly add information
What we learned
Not only was mindeye our team's first ever facial recognition project, it was also our first ever cloud vision project in general. Learning about how the nuances of OpenCV worked, though difficult, proved to be incredibly rewarding in the end, as all out teammates have now developed a much deeper sense of how the software works. We look forward to applying what we learned in future projects nad hackathons.
What's next for mindeye
We believe that mindeye has a lot of future growth and potential. The effects of Alzheimer's disease are long and far reaching; they do not merely encompass the victim's inability to recall past relationships and recognize faces. Victims also experience inability to recognize common objects, an inability to think clearly, and difficulty concentrating. As a result they may become agitated easily or may experience a personality change. We plan to implement object recognition one day to assist with the victim's ability to recognize everyday objects. In the future, we hope to be able to implement mindeye as a wearable, similar to Google glass. This way, the victims of Alzheimer's disease are going to be able to constantly have the reminders provided by mindeye.