As a freshman, our team member Janie was diagnosed with encephalitis, a neurological illness that gave her such heavy "brain fog" that she wasn't able to immediately recognize her own closest friends and family. There are many other medical conditions that prevent people from recognizing the most import people in their lives, including Alzheimers, dementia, and visual impairments. We wanted to build a device that could simplify the lives of those affected by these conditions by discretely identifying known persons.
What it does
Our device has three key functions: 1) facial recognition 2) medication reminders 3) patient analytics for healthcare providers
How we built it
We connected a raspberry pi to a webcam which, at the push of a button, transmits image data to our facial recognition API. From this API, the patient receives either the name of the person in the image, or a notification that no familiar faces were identified. We are using IBM Watson to communicate image results, regular medication reminders, and other messages.
Challenges we ran into
One major challenge was packaging all this functionality into a discreet device. If we had access to additional resources, we would 3D print a custom body and use smaller components to make the device less bulky and eye-catching.
Another big challenge was getting facial recognition working. We tried multiple APIs in addition to running facial recognition software on-device and in the cloud, and after a day for searching for one that actually worked, we finally found one that we could build off of.
Using IBM Watson, it was challenging to use figure out the right way to analyze all the words a person speaks in order to get useful data about their disorder to send to a caregiver. For example, demensia
What's next for memoreyes.org
Making it better: 1) A sleeker, more discreet design 2) QR code scanning to verify that the pill pack the patient has in hand is meant to be taken at the current time 3) Directional vibration alerts for incoming objects