What it does
ElderVision will alert your family if your elder falls. When a picture matches the trained model of the Watson API, an event will trigger, sending a MMS to all the registered family members. With the received image, the family members will then judge if it's a false alert or if they have to intervene.
How we built it
We built a simple Android app that is sending data through a HTTP POST to the API that we built in C#. The API then manages the calls to the IBM Watson API and send a MMS if it's applicable.
Challenges we ran into
- We had problems with compressing the images to send to the API.
- We also had problems to train our model with IBM Watson API.
Accomplishments that we're proud of
- We're proud of getting the IBM Watson API to work and to return a pretty accurate result when someone is falling.
What we learned
We learned that taking the time to start with a good plan can help a lot through all the process of making the solution. When we drift from the plan, we remind us of the plan and change it, if needed.
What's next for ElderVision
The next step is to implement a video feed on a camera of our own (e.g. with a RaspPi) to reduce the cost and increase the flexibility.