BetaBeta: works today, with accuracy limits explained below.

Vehicle colour and type detection on the CCTV you already have

When something happens, the first question is usually "what did it look like?" Eye AI adds descriptive details to what your cameras already see. For each vehicle it estimates type and colour. For each person it estimates clothing colour and whether they carry a bag, umbrella or suitcase. These details make events searchable, so you find the red van or the person with a bag in minutes, not hours.

ILLUSTRATION WITH SAMPLE DATA

The problem

Looking for "a white car, around lunchtime" means watching recordings for hours.

Incident reports say "a dark vehicle" because nobody can search by detail.

Gate logs record that something passed, but not what it was.

How it works

  1. Camera check first.In the free camera-readiness check we review your views and say which are close and clear enough for details to be useful.
  2. Details are added automatically.As people and vehicles are seen, Eye AI attaches its best estimate of type and colour to each one.
  3. Analysis stays on site.A computer at your site does the work. Your video is not sent away.
  4. Search by detail.On your password-protected dashboard you filter events by vehicle colour, vehicle type, clothing colour or bag, each with a snapshot.

What you get

  • Vehicle colour and type on each logged vehicle.
  • Clothing colours for people.
  • Flags for bags, umbrellas and suitcases.
  • Filtering by these details in the searchable event history, with CSV export.
  • Details attached to counts, parking visits and gate events, so reports say more than "one vehicle".

Where it is used

Honest limits

  • This analytic is in beta, and everything it reports is an estimate, not a fact.
  • Vehicle colour is far more reliable than clothing colour. Clothing colour is a rough estimate and should be used to narrow a search, not to confirm who someone is.
  • Coloured light, night-time, glare, shadows and wet or dirty lenses change how colours look and can cause mislabelling.
  • Small, distant or partly hidden people and vehicles may get no details at all.
  • Eye AI does not identify who a person is, and we do not state age as fact. Details describe what is seen; they are not a profile of anyone.
  • We publish no accuracy figure for this beta.

Everything Eye AI cannot do →

Questions about vehicle and person details

Does it work with my current cameras?

Yes, when they provide a standard video stream and the view is close and clear enough. We check this first, free of charge.

How reliable is clothing colour?

It is a rough estimate. Use it to narrow a search, then check the snapshot yourself.

Can it tell me who a person is?

No. It describes what is visible. It does not identify people.

Does it work at night?

Less well. Low light and coloured lighting reduce reliability, especially for clothing colour.

Is my footage uploaded anywhere?

No. It is analysed on a computer at your site.

Tell us about your site.

We read every request and reply within 24–48 hours (working days, Monday to Friday) with questions and a proposed solution.

Request a solution
Request a solution