Inspiration
We thought about how long it takes to sign people into dormitories and wished there was a process where we could just scan our IDs and automatically sign them in.
What it does
Basically, it turns a regular laptop and webcam into a smart check-in kiosk. A dorm host steps up, holds their UMBC ID to the camera, and the system logs them in. From there, the system tracks the host's ID and ties it directly to whoever they want to sign in as a guest.
Challenges we ran into
Getting all these heavy AI libraries to run smoothly on a local Windows machine was a massive headache. Fighting with C++ build tools just to get dlib and face_recognition installed took way more time than expected. Another huge challenge was lag. Running a text-reading AI and face detection on every single frame of a live video feed will completely freeze a desktop app. We had to build a "frame skip" system so the heavy AI calculations only run every 5 frames, keeping the Tkinter GUI and camera feed buttery smooth.
Accomplishments that we're proud of
We are mostly proud of learning a completely new thing (computer vision) and working hard to figure out a solution to a real problem. Getting the main core process of the project working smoothly was a huge win for us.
What we learned
We learned how to connect raw computer vision to a clickable user interface that actually saves and manages data. We also learned that AI is only as good as the image you feed it—forcing users to align their ID inside on-screen "bounding boxes" saved us from writing hundreds of lines of error-handling code.
What's next for UMBC ID Checker
We want to add better support for different formats of standard driver's licenses and implement a true facial recognition process to match the live person standing at the desk to the photo printed on the ID.
Log in or sign up for Devpost to join the conversation.