We were inspired by the modern problem of having safe and secure physical storage but without the hassle of having to have a password memorized.
What it does
Our app takes a file and encrypts it or decrypts it and allows the user to select a location to store the resulting file. When a file is selected, the facial recognition system is triggered and either takes a photo of the user to add to the dataset or it will compare the current user's face with existing photos in the database and unlock a file if there is a match. We also used different APIs to track a user's location and send them an email containing login details. We also implemented a failsafe in the form of a QR code that you can present to the camera to unlock a file in the case that the facial recognition system doesn't work.
How we built it
We used python as well as several different APIs and modules to allow us to perform a variety of tasks. We used OpenCV for our facial recognition system. We created a dataset of different users and photos associated with them to train the supervised machine learning algorithm. We used pyAesCrypt to encrypt and decrypt files. We used geolocation, google maps's API, to track the location of the user when a login is attempted and the details are emailed to the user.
Challenges we ran into
We faced a large variety of challenges. One problem we had was getting the facial recognition system to be accurate as well as having it easily distinguish different users. The encryption system was also challenging to implement due to the file type extensions.
Accomplishments that we're proud of
Throughout this hackathon we actively learned about different algorithms and their implementations through trial and error and by attending workshops for AI.