The idea.
Can vehicle detections help estimate which parking spaces are available?
Computer Vision · 2026
ParkEye
Finding open parking through computer vision.
THE PROJECT
Hackathon prototype · March 2026
MY CONTRIBUTION
Computer vision prototype (five-person team)
The work.
With a five-person team, I worked on YOLOv8s confidence and IoU thresholds and reviewed annotated frames in FiftyOne.
The result.
A working demo built and presented in a 35-minute hackathon sprint.
How it works
I worked in a five-person team to build and demonstrate the proof of concept. We adjusted YOLOv8s confidence and IoU thresholds and inspected annotated frames in FiftyOne to understand the detector’s behavior.
Tools Python · YOLOv8 · OpenCV · FiftyOne
- Used Python and OpenCV around the vehicle-detection pipeline.
- Adjusted detection confidence and IoU thresholds rather than treating every prediction as ground truth.
- Reviewed annotated frames in FiftyOne to make outputs inspectable during the sprint.
Camera frames → OpenCV → YOLOv8s → Vacancy estimate