Our inspiration is from the smart home IoT control that Mark Zuckerberg is building recently. We want to make the entry process to a home or office to be remotely accessible and secure.
What it does
Our knock-knock project is able to recognize people at the front door and send the image to the home owner's phone. By replying to the message, the owner will be able to decide if the person should come in and if he wants to save to person's information into the database for future facial-recognition.
How we built it
We built the face recognition technology using Microsofts Face API. This API included features such as recognizing faces and generating ids. Previously generated id's were used to compare with the current face id's to see the similarities. We also built the text messaging base using Twilio API. MMS and SMS functions were used.
Challenges we ran into
The most challenging part of every hackathon is the time constraint. We had to ensure that our project was complete in 2 days with thorough testing. We learned a lot about python post requests and used a lot of different API's that we hadn't used before that added to the difficulty of the task. We did not have a monitor available and had to run the raspberry pi 'headless', which meant using vnc viewer and ssh to run commands and make changes on the pi. Opencv took a long time to compile on the raspberry pi. Also many videoCapture functions don't work on mac system, which took us some time to fix.
Accomplishments that we're proud of
We are very excited about the work we have done in general. We are proud of stepping into the world of smart IoT devices through this project. We are also proud of the way we used multiple API's from multiple sources and brought them together through our implementations. We think that this product could be a cheap and effective solution to increase household security especially in areas that have higher crime rates than usual. Also, if the owner of the house or anyone who is living in the house is ever locked out this technology will remotely open the door through our commands. We think that having such opportunity through a simple idea is something to be proud of.
What we learned
Through this process we learned the basics of face recognition and we started exploring the world of home security.
What's next for Knock-knock
Next we are thinking to add the a sensor on the door so that it could tell us how far the person is from the front door. If the distance is in a certain range, our "knock knock" program will run automatically. This will save a lot of energy.