MECHAE

Inspiration

Every 19 minutes, one elder (65+ years) dies from falling down. Falls are the leading cause of emergency room visits among the elderly, racking up 2.5 million visits annually. It's a $80 billion space, and the market is flooded with wearable equipment that have expensive recurring fees.

What it does

Our detection algorithms run on the cameras installed in open spaces, particularly in elderly homes, and fires up alerts once a fall has been detected. The patient is recognised from the model and contact information of emergency contact is queried from the database and the emergency contacts are alerted by text.

How I built it

We tried Clarif.ai and did not get the accuracy we wanted for our custom models. Then we went to the Google Talk and realized we had the option for AutoML, which turned out to be a winner.

Challenges I ran into

Dealing with the Cloud AutoML API

Accomplishments that I'm proud of

Integrating the entire tech pipeline to the Cloud AutoML. Our team participated in most of the events and we got a ton load of stickers!

What I learned

Teamwork, AutoML APIs

What's next for MECHAE:

Increasing the scope of the project to prevent swimming pool fatalities, fires and security scares.

Built With

• docker
• flask
• google-cloud
• google-cloud-kubenertes
• google-cloud-vision
• mongodb
• ngnix
• uwsgi