Inspiration
I have always been inspired by blockchains and encryption, and especially love finding ways to make data more secure.
What it does
Our app uses machine learning to identify an object and assign a crypto key to it by taking a picture. Pictures of this object can then be used to decrypt files encrypted using the picture of the object.
How we built it
We used mlkit pre-trained visual tags for android studio with Java to develop a program to identify objects, and we used AES encryption to encrypt the data using the key generated for the picture.
Challenges we ran into
I originally planned on connecting the app to a computer to encrypt files on the computer, but I decided to start this project very late last night and ran out of time. I ended up encrypting local files on the android device. Object recognition was also really hard and not super secure yet since I didn't train the AI on the specific object, and instead generate the key from the color of the object and the name of the type of object.
Accomplishments that we're proud of
I am very proud of the visual tag system that identifies different objects.
What we learned
I learned how to use the MLKit API for java with android studio to use pre-trained machine learning models.
What's next for Pictocrypt
Pictocrypt will become more secure in the future and will have more advanced object recognition software, as well as RSA encryption.
Log in or sign up for Devpost to join the conversation.