The problem
Bring real-time computer vision to constrained hardware, and package it cleanly enough that a classroom of students could replicate it.
What I built
A complete on-device vision build for the Raspberry Pi 5 with an accelerator camera HAT: a Bash setup script that provisions the device and dependencies, a Python detection module that runs trained vision models against a live camera feed, and a curated models directory of deployed network weights.
94.51%
live detection confidence
The outcome
A paid project through Forsyth Technical Community College’s IT Club, built by a three-person team. Taught the full build to a class of four students. The work will be published by the college.
Interested in this kind of work? Get in touch