We saw the current smart locks that are available for sale and thought that they were cumbersome and dificult to use.
What it does
Using a Kinect, the lock is able to take a picture of who is coming to the door and automatically unlock the door if the person is registered in the system. If not, the owner is alerted and can have a video call with whoever is at the door.
How we built it
We used the Kinect SDK to program the Kinect to take a picture of the person at the door and send a request to a Nodejs server that processes the request and handles video chat and then sends the facial recognition data to a python service that will then match the picture to the registered users,
Challenges we ran into
We had a lot of technology that we tried to incorporate into the project but were not able to. For example, we tried to use a Qualcomm Dragonboard but we couldn't find a way to connect sensors to the board.
Accomplishments that we're proud of
We were able to get the face recognition and machine learning done.
What we learned
We learned that we needed to manage our time better. We spent too much time setting up the hardware which mean we had to rush the code to a certain extent.
What's next for Let Me In
We would like to make it more accurate by incorporating different technologies like a 3d depth scanner.