An interactive computer vision installation that transforms human presence into generative visual environments. Built with Unity and a custom multi-camera detection server.

Computer Vision YOLOv8 FastAPI HTTP Polling Unity SpeedTree

Input

Networked webcams observe high-traffic zones and turn occupancy into smoothed per-camera metrics.

Translation

A Python detection server uses OpenCV, YOLOv8, and FastAPI; Unity polls the metrics endpoint over HTTP.

Output

Unity micro-zones grow, decay, breathe, and trigger environmental events from live building activity.

Multi-camera person detection dashboard
Scaled detection system with 3 cameras
Unity interface with Data Forest scene controls