re.mind

Inspiration

Dementia is a major issue in today's society that slowly erodes the memories of the elderly and their abilities to recognize even the most familiar faces. To combat this issue we developed re.mind, an application that recognizes and gently nudges the patient towards remembering their precious memories

What it does

Re.mind is an application we developed that is constantly scanning the user's environment until it recognizes an individual that is related to the user. Once this is done it allows the user to ask questions regarding that individual such as their name, their favorite memory and other such information. Doing research on the issue of dementia also led us to cognitive stimulation therapy which we attempted to implement into our application by gently guiding the user to an answer instead of directly giving it to them.

How we built it

We built re.mind primarily using android studio and Kotlin but we used many other APIs and frameworks to get the final product. We used google's MLKit in tandem with TensorFlow Lite for the facial recognition aspect of the app and used Google's Gemini API to generate the responses to the user's queries.

Challenges we ran into

Some of our biggest challenges were dealing with the lack of sleep and difficulty in implementing the harder parts of the project. Parts that dealt with the facial recognition especially were difficult when it came to understanding the image cropping and how to generate vector maps of the faces to compare them. We also had a lot of git issues that took a while to resolve.

Accomplishments that we're proud of

Finishing the application in such a limited amount of time is what we as a team are most proud of. Overcoming challenges and hurdles along the way to complete a project that is mostly fleshed out and learning new tech stacks like Kotlin and TensorFlow while coming in with a limited understanding of how they worked also allowed us to learn a lot during this project which we are very proud of.

What we learned

We learned how to work together while emotions are running high and how to collaborate more efficiently for future hackathons. We learned how to develop apps using Android Studio and how to integrate libraries and APIs with our application.

What's next for re.mind

We plan to expand re.mind's functionality to include being able to identify every day objects and provide an explanation on how to use them for extreme cases of dementia and enable the user to add more specific information for each relationship that they have. Our end goal is to have the app act as a virtual caretaker to reduce the financial burdens that come with having to hire a human caretaker while retaining much of the same functionality.

Built With

• kotlin
• tensorflow