The inspiration from this project came from a conversation had between all the group members where we were talking about the growing crime rates in our areas and how we wish we had something that could help us feel safer while at home.
What it does
This application recognizes faces in order to notify the owner if there is a known or unknown person that is approaching their residence. This system helps the user safeguard their house while they are home or away.
How we built it
We built and trained the facial recognition model using python, OpenCV, a standard web camera, and a Raspberry Pi. We integrated this with a front-end mobile application built using Android Studios in Java. The database was implemented using Google cloud's Firebase Database and Storage.
Challenges we ran into
Some challenges we ran into include spending a tremendous amount of time training the data and compiling the various libraries. We also struggled to set up the storage and database. We also considerably struggled on sending and receiving notifications for the android application.
Accomplishments that we're proud of
Overall, we are extremely proud of this project because of how little we knew about this project going in, and how much we learned coming out. Out of all 4 members of our group, not a single person had ever used OpenCV, trained a machine learning database, built an Android app, or utilized a real-time database. We are extremely proud of the progress we made in all of these fields, and the fact that we were able to get a viable working project out by the deadline.
What we learned
Overall, we learned a lot about machine learning, android app development, databases, front and back integrations, and facial recognition.
What's next for Facial Recognition Security Camera
We would love to add real-time streaming, support for multiple cameras, and motion-detecting sensors.