María

• 

  A picture of the assembled hardware for the device.

• 

  An example of the computer version, fully-functioning.

• 

  A second example of the computer version, fully-functioning.

Inspiration

Our inspiration was the inability of some dementia patients to even recognize their own family members. Izzie's grandpa had dementia, and he would at time mistake his wife for his sister, or forget his grandchildren. With that in mind, we wanted to create a device that would remind a dementia patient of who the person is that they are talking to.

What it does

There are two code versions on the repository. One is a computer version that listens to 7 seconds of audio, where a person would say their name. It then converts that speech to text using rev.ai, and pulls up the person's name, their relationship with the dementia patient, and a picture of them together to jog the memory and recall previous emotions. The second version is for the hardware. It has most of the same functionality, but the speech input was not completed, so it currently runs using preinstalled sound files.

How we built it

We programmed the raspberry pi with python via SSH connection. We also connected an arduino for the purpose of analog-to-digital conversion as we attempted to add the sound input to our hardware.

Challenges we ran into

We ran into a lot of challenges with installing the necessary python packages, especially with numpy. Configuration issues with the raspberry pi we were using meant we spent hours trying to uninstall and reinstall packages instead of moving forward with our code.

Accomplishments that we're proud of

We are proud of having a finished product as this was each member's first time at a hackathon.

What we learned

We learned how to use APIs and extended our knowledge of linux and command prompt.

What's next for María

The next step would to get the data from the onboard speaker transporting correctly to the raspberry pi. More importantly, however, we would need to create an easy way for users to upload text with information about the person in question and relevant photos to the program. In addition, we’d like to use a feature similar to “hey Siri” or “Alexa” works in that the device listens for someone’s name after you say “my name is” and pulls up the corresponding information to the screen.

Built With

• python
• rev.ai