Computer Vision

Real-time visual intelligence from your existing cameras.

Computer Vision analyses video from your existing cameras to detect safety violations, product defects, security events and attendance. For each event, your team receives a clip and the reason it was raised, rather than hours of footage to review.

Last updated 9 November 2026

Book a demoAll products
Live feedCortvu Vision
PersonHelmet ✓No helmetDefectCAM 01
  • Worker detected in Zone A09:41:07
  • Safety helmet check passed09:41:07
  • No helmet detected. Supervisor alerted09:41:12
  • Defect on line 2 flagged for review09:41:20

Illustration. Hover an event to locate it in the image.

Capabilities

What does Computer Vision deliver?

People and object detection

Counts and tracks people, vehicles and objects within defined zones.

Safety compliance

Detects missing helmets, vests and other protective equipment in real time.

Defect detection

Identifies damaged or non-conforming products on the line and flags them for review.

Security alerts

Raises alerts for after-hours movement, forced doors and restricted areas, with video evidence and context. See the Security section.

Counting and queue analytics

Footfall, line counts and queue length for retail and operations teams.

Face recognition attendance

Recognises staff at the entrance and records attendance. See the Attendance section.

Implementation

How is Computer Vision set up?

  1. 1STEP 01

    Connect cameras

    Your existing IP cameras are connected on site or over your network.

  2. 2STEP 02

    Define requirements

    We agree the zones, rules and events that matter. A focused scope is recommended at the start.

  3. 3STEP 03

    Monitor

    Detections are delivered as alerts with a clip and the reason they were raised.

  4. 4STEP 04

    Refine

    Reviewer feedback is used to tune the rules and reduce false alerts during the pilot.

Face recognition attendance

How does face recognition attendance work?

Face recognition attendance uses the existing entrance camera to recognise enrolled staff as they arrive and records the time automatically. Staff give consent first, and HR reviews exceptions such as late arrivals or missed scans instead of keeping manual registers.

  • Contactless check-in

    No cards, PINs or fingerprints are required. Staff walk through the entrance as normal.

  • Accurate time records

    Arrival and departure are recorded automatically, with late arrival, early departure and overtime flagged.

  • Payroll-ready reporting

    Daily and monthly reports are exported as CSV or sent to your HR or payroll system.

  • Multi-site visibility

    A consolidated view across offices, factories and sites, including visitor records.

✓ Entry logged 08:47OFFICEExisting CCTVAyesha K. ✓Main entrance · illustrative

A person approaches the office entrance

Illustration with sample names, not real data.

  1. 1

    Consent and enrolment

    Each employee gives consent and completes enrolment. Employees who decline use an alternative method.

  2. 2

    Recognition at entry

    The entrance camera recognises enrolled staff on arrival, and the door can be released automatically.

  3. 3

    Exception review

    HR reviews mismatches, missed scans and late arrivals in a single list.

Consent and privacy

Face data is biometric data and is regulated in many jurisdictions, for example under GDPR in the European Union and BIPA in Illinois. We implement consent, retention and access controls with you. Please obtain legal advice for your location.

  • Enrolment requires employee consent
  • Faces are stored as encrypted templates, not photographs
  • Retention periods are defined by you, and data is deleted on request
  • A person reviews any result the system is unsure about
  • All access is recorded in an audit log
Book an attendance demoWorks with your existing entrance camera or a tablet at the door.

Security monitoring

How does Cortvu reduce false security alarms?

Cortvu checks the time, the location and who is authorised to be present before it raises an alert, then sends the video clip and the reason to the right person. This is how it cuts false alarms and speeds up the response.

  • Contextual alerts

    Each alert includes a clip and a clear statement of why it was raised.

  • Access monitoring

    Track who enters each area. Unrecognised people and after-hours entry are flagged.

  • Time-based rules

    Rules vary by time of day and can be set per zone and schedule.

  • Doors and perimeters

    Detects forced doors, doors left open and movement along a perimeter.

  • Visitor records

    Visitors are recorded alongside staff and are searchable by date and area.

  • Incident timeline

    All events and actions are consolidated in one timeline for review and reporting.

Alert inboxSample data
4 raised1 auto-resolved3 need action
Reason for alert

After hours, no badge scan, person not recognised

Checks performed
  • Schedule: closed 19:00 to 07:00
  • Door log: no badge entry
  • Camera 04, 12 s clip

Illustration with sample data. The buttons do not send anything.

Access to footage

Footage is sensitive. You control who may view it and how long it is retained, and every view is logged.

  • Only authorised users can view footage
  • Every view, export and action is recorded in an audit log
  • Footage is deleted automatically after the retention period you set
  • The system supports your security team and does not replace it
  • Rules and schedules can be changed at any time
Book a security demoWorks with your existing cameras. Door and alarm integrations are agreed with you.

Applications

Who uses Computer Vision?

  • Manufacturing
  • Warehouses
  • Retail stores
  • Offices and facilities
  • Construction
  • Multi-site organisations

Requirements

What do you need to start?

  • IP cameras that provide a standard video stream
  • A network connection, or a small device installed at your site
  • A defined list of events to detect
  • A designated person to review alerts

FAQ

Questions about Computer Vision

Ask us in a demo
Is new camera hardware required?

In most cases, no. Standard IP cameras are supported. We assess each camera during the initial assessment and advise if one is unsuitable because of age, lighting or placement.

Does video leave our premises?

It does not have to. Detection can run on a device at your site. You decide what is retained and for how long.

How long does a pilot take?

It depends on camera access and how clearly the use case is defined. We agree a timeline with you in the written proposal.

What happens when the system is unsure?

The case is routed to a person for review. Uncertain detections do not trigger alerts or actions on their own.

How are false alarms reduced?

Before an alert is raised, the system evaluates the time, the zone and who is authorised to be there. Rules are then tuned with your feedback during the pilot. Performance improves over the pilot and is not perfect from the first day.

Does it replace the security team?

No. It provides the team with better-qualified alerts. People remain responsible for the response.

Can it integrate with our door or alarm system?

In many cases, yes. It depends on the system, and the integration is assessed together at the start.

Is face recognition attendance permitted?

It depends on your jurisdiction. Biometric data is regulated in many regions, for example under GDPR in the European Union and BIPA in Illinois. We implement consent and retention controls with you. Please obtain legal advice for your location.

Can employees decline face recognition?

They should be able to. An alternative method, such as a card, a PIN or sign-in with HR, is provided for anyone who declines.

More from Cortvu

Related products

AI Agents

Workflow automation and conversational AI

View AI Agents

See Computer Vision on your own cameras

Book a free demo. We will demonstrate the system in your own environment.

Book a demo