Investigations

Presence

Python February 2026

A computer vision welfare monitoring system using Python for detecting activity patterns and alerting on anomalies.

Presence is a welfare monitoring system that uses computer vision to detect activity patterns. The idea is simple: cameras in a care environment detect normal activity patterns, and the system alerts when those patterns change — someone hasn't moved in an unusual amount of time, or activity levels drop below baseline.

The Python backend processes video feeds, extracts motion data, builds activity profiles, and compares real-time behaviour against established baselines. Alerts trigger when deviations exceed configurable thresholds.

It's a sensitive application domain, so the focus was on minimising false positives while ensuring genuine anomalies get flagged quickly. The system learns what "normal" looks like for each monitored space.

What this project demonstrates

Computer vision activity pattern detection
Baseline learning for normal behaviour profiles
Configurable alert thresholds for anomaly detection
Real-time video feed processing
Designed for care environment monitoring
Focus on minimising false positives

Technologies

Computer VisionPythonMonitoringHealthcare

© 2026 Jonathan Leahy · v1.0.1