Block 4. Video Analytics

1

The Video-Analytics Map: What a Surveillance System Can Detect

A field guide to video analytics software: the seven analytics a surveillance system can detect, where each runs, and the honest accuracy of each.

2

Object Detection and Classification in Surveillance

How object detection finds people and vehicles in a camera scene, turns them into searchable VMS events, and cuts false alarms — with honest accuracy.

3

Object Tracking and Re-Identification Across Cameras

How object tracking and re-identification follow a person or vehicle across cameras, how accurate it really is, and why a movement trail is personal data.

4

Face Recognition in Surveillance: How It Works and Where It Is Restricted

How face recognition works in video surveillance, how accurate it really is outside the lab, and where GDPR, the EU AI Act, and BIPA restrict it.

5

License-Plate Recognition (LPR / ANPR)

How license-plate recognition (LPR/ANPR) works, why its accuracy is capture rate times read rate, and why a plate is personal data but not biometric.

6

Behavioral Analytics: Loitering, Intrusion, Crowd, and Zones

How behavioral video analytics works — line crossing, intrusion zones, loitering, crowd counting — why false alarms are the metric, and the privacy line.

7

Anomaly Detection in Surveillance Video

How anomaly detection in surveillance video flags the unusual with no rule, why its accuracy is a false-alarm tradeoff, where it runs, and its limits.

8

Search by Event: Making Months of Footage Findable

How forensic search by event turns months of surveillance footage into a searchable index, what you can query, the ONVIF limit, and the privacy line.

9

Tuning Analytics: False Alarms, Accuracy, and the Operator's Reality

Why video analytics accuracy is a range, not a number: the precision/recall dial, the false-alarm filter cascade, the base-rate trap, and alert fatigue.