Simat Labs

ICA is coming soon

ICA — In-Class Analysis

AI SDKv.1.13

ICA — In-Class Analysis

Real-time behaviour analysis for schools. Detects attention, hand raises, and incidents across classroom cameras.

ICA (In-Class Analysis) is a real-time behaviour analysis solution that monitors student attention, hand raises, and classroom incidents across multiple RTSP cameras simultaneously.
It gives school administrators and teachers instant visibility into classroom dynamics — surfacing disengagement, participation, and incidents as they happen, without manual review.

ICA runs on a central on-premise server using a single GPU to batch inference across all streams. No video or biometric data leaves the school network.Behaviour detection covers attention levels, hand raises, and disruptive incidents including fighting, eating at the desk, and potential cheating — with configurable thresholds per camera.

Multi-camera

Processes all classroom RTSP streams on a single on-premise GPU. Scales from one room to an entire school.

Real-time alerts

WebSocket alerts surface hand raises, low attention, and incidents the moment they occur — no manual review.

Privacy-first

On-premise deployment, explicit consent flows for minors, and no biometric data uploaded to the cloud.

Frequently asked

What behaviours does ICA detect?

ICA detects a range of student behaviours in real time: attention levels (looking at board vs. distracted), hand raises, and disruptive incidents including fighting, eating at the desk, and potential cheating. Each behaviour category has configurable sensitivity thresholds set by the school administrator.

What hardware does ICA require?

ICA runs on your existing RTSP cameras — no new cameras required for most schools. A central on-premise server with a single GPU handles inference across all streams. Simat Labs specifies compatible server configurations during pilot scoping, and typical deployments go live within two to three weeks of server arrival.

How does ICA protect student privacy?

All video processing and face recognition happen on the school's own server — no footage or biometric data leaves the building. ICA is designed with explicit consent flows for minors from day one, supports the right to erasure, and aligns with GDPR principles including purpose limitation and data minimisation. Actual compliance depends on school configuration and local law; Simat Labs works alongside your legal and safeguarding teams during rollout.

How are alerts delivered?

ICA streams events over WebSocket to a real-time dashboard visible to teachers and administrators. Alert types and thresholds are configurable per room, per behaviour, and per time window. Alerts can also be forwarded to third-party systems via a local API — no internet connection required at alert delivery time.

Last updated: