https://devpost.com/software/elderguardian
Authors: Ryan Fusco Nate Kupec Bradley Zaricki
We were inspired through personal experiences to create a protective software for elderly people on hospice or in retirement homes.
ElderGuardian is a discreet in-room device that continuously monitors conversations. It can distinguish between different voices and assign them to individual profiles. If concerning behavior or malpractice occurs, loved ones gain secure dashboard access to view transcribed recordings, complete with timestamps and speaker identification.
ElderGuardian was built using MongoDB, Flask, and React.
The speaker diarization was very tricky to implement, as well as the language classifier and the dashboard.
Our database is fully connected to our frontend providing a seamless flow of information from recieved audio to the frontend dashboard.
We learned about audio analysis, planning out design flows, and database tools.
The next step will be to implement a continous voice monitoring device to provide a full, shippable product.
Credit: https://github.com/MahmoudAshraf97/whisper-diarization