AI video analytics

SentinelVue

Turn a camera network into real-time alerts and answers.

SentinelVue analyses live and recorded video to detect and track people, vehicles, objects and number plates across many cameras at once. It raises alerts for security events, counts people and vehicles without double counting, and gives each user only the cameras and alerts their role allows.

Interactive demo

Try it here.

Click through the screens. Everything you see is sample data. For a live demo on your own data, book one with us.

SentinelVueDemo data

The problem

Hundreds of cameras, and nobody can watch them all.

  • Operators miss events on busy video walls.
  • Finding a person or vehicle in recorded footage takes hours.
  • Basic motion alerts produce too many false alarms.
  • Visitor and traffic counts are guesses.

How it works

From connect cameras to alert and report.

  1. 01Connect cameras

    Live streams or archived footage from one site or thousands of cameras.

  2. 02Detect

    People, faces, vehicles, number plates and objects in every frame.

  3. 03Match and track

    Against watchlists and across cameras, in under a second.

  4. 04Alert and report

    Role-based alerts, searchable events and counting reports.

Features

What it does.

Multi-object analytics

Recognises people, vehicles and objects at the same time, in live streams or archived footage.

Fast, robust matching

Face detection and database matching in under a second, even with masks, glasses, headgear, ageing, low light, partial views or crowds.

Number-plate recognition

Vehicles and plates read and searchable by time and camera.

Anti-spoofing and liveness

Tells a real person from a photo or screen. Available as an optional plug-in.

Counting without duplicates

People, body and vehicle counts across several cameras, without counting the same person twice.

Visitor analytics

Footfall trends, occupancy, unique and returning visitors, and crowd density for malls, stations and venues.

Role-based access

Administrator, supervisor and watcher roles, camera groups and watchlist permissions.

Tuned notifications and audit

Each team sees only the alerts it needs. Every user action is logged.

How we set it up in Japan

Face recognition is configured only for a purpose the customer defines and publishes, with notices at the site, retention limits and access logging, following Japan's Act on the Protection of Personal Information. Counting and vehicle analytics can run without storing faces at all.

Deployment

  • On-premise server
  • Edge / offline

Security

  • Role-based access by camera group and watchlist
  • Complete audit log of user activity
  • Configurable retention
  • Runs on your own servers

Integrations

  • IP cameras (RTSP / ONVIF)
  • Video management systems
  • Access-control systems
  • REST API and webhooks

Who uses it

Built for.

  • City and public-space safety
  • Transport hubs and stations
  • Malls, stadiums and venues (counting)
  • Corporate campuses and gates
  • Retail loss prevention

See SentinelVue on your own use case.

We'll arrange a live demo with our partner's engineers, online or at your office, in Japanese or English.