Verifex

VIDEO https://drive.google.com/file/d/1asTUI5vQrn_ESTuqB9IG7v1PPP4fJbku/view?usp=sharing

Inspiration

During the school year, I had heard about students asking other students to write essays for them - some for money and others for free. This was a significant issue that even teachers were concerned about, holding an assembly regarding academic honesty/dishonesty.

What it does

Verifex takes information from professors/teachers and creates a series of 5 questions for the user to respond to. It uses continuous facial verification to ensure that the person in front of the webcam, answering the questions, is in fact the student. It also records the amount of time taken, ensuring that a student does not take so long that they can search for the text itself (not presented in the video).

How we built it

I built the application using React. ChakraUI was used for components. I used react-native-webcam to record the user. GunDB, a peer-to-peer storage framework, was used to store the data in a decentralized manner. I used FastAPI and Deepface to load and use models for facial verification, and to make the facial verification available to the user.

Challenges we ran into

An initial implementation ran into some trouble using a browser-based facial verification system. However, the model took a lot of time to load and was freezing up the screen. Therefore, I decided to move last-minute to a server based model for the time being.

Accomplishments that we're proud of

Starting yesterday, I wasn't sure that I was going be able to finish, but I was surprised that I was able to solve a topical problem, present a good UI, and have solid functionality in such short time!