Questions · Cameras
Why can you see that somebody came in but not recognise their face?
Because detecting and identifying are not the same task and do not need the same detail. With very few pixels on a person you can already see there is somebody and which way they are going; to recognise someone you already know takes several times more, and to identify a stranger —to put a name to a face you had not seen— takes about ten times the pixels of detection on that same person.
The standard that sets this out is UNE-EN 62676-4, the application guidelines for video surveillance systems. It defines five levels in pixels per metre of scene: monitoring the scene, detecting that somebody is there, observing what they are doing, recognising someone you already know and identifying a stranger, and it gives each one a figure that climbs in big steps up to the two hundred and fifty pixels per metre for identification. It lets you talk in numbers at the design stage instead of in adjectives.
The important part is that this level is set point by point and not for the whole installation. The same camera can be identifying at a metre and a half from the turnstile and only detecting at twenty metres, and both are right if they were decided beforehand. The problem turns up when nobody decided it and you find out on the day of the incident.
Where it is set out in full.
All the questions.
Five hundred and forty-four, grouped by topic. Each with its own page.
Get started
See it working.
A written answer is enough to decide whether it interests you. To know whether it works at your site you have to see it on a real scene, with the image as it is and, on top of it, what the system understands from it. That is on the IRIS Neural website, which is the group's product.