What is our inspiration

We have family members that are suffering from early symptoms of Alzheimer’s and we felt motivated to improve their quality of life especially during a difficult diagnosis. The concept of notifying the user of people's identities was partly inspired by doorbell apps that releases a notification and activates when it detects movement.

What it does

The program detects a person’s identity through facial recognition using the camera and outputs an image accompanied by a brief description that shows a significant event in their past. By invoking strong emotions within the patient, we intend to support their long-term memory and maintain their relationships with loved ones.

How we built it

We mainly used the custom tkinter and tkinter GUI library to build the user interface and the “deepface” model from the OpenCV library to develop facial recognition for our program.

Challenges we ran into

Since we were new to customtkinter, a significant amount of time was spent on learning the GUI library, which delayed the development of the GUI and limited its functionality and design. Connecting the input from the camera to the user interface was the most challenging as we did not have any experience with integrating facial recognition into a GUI program.

Accomplishments that we're proud of

Our biggest accomplishment is having our main program function correctly and being able to transfer data from the camera to the GUI to output the information when the facial recognition detects a person. We were also able to create and functional GUI despite our lack of experience with the custom tkinter library.

What we learned

We learned many aspects on how memory works during our initial research. We also recognized that it is important for Alzeimer’s patients to have a way to help remember their loved ones without feeling left out. A crucial lesson we learned is also to be more prepared for hackathons, such as learning GUI libraries beforehand to save some time.

What's next for The Memory Hub

If we continued working on this project, we would implement the rest of our features, such as the ability to add or remove data from the app. Instead of hardcoding in the output, we would make the program search through a set of data to find the person it detected and display the appropriate information. We could also make the program randomly pick a memory for the same individual if there are multiple entries in the data.

Built With

Share this project:

Updates