AdEye Concept presentation · July 2026

Street video ads. Finally measurable. Finally personal.

  • Track street impressions Online ads know who saw them. Street screens never did. Now they can.
  • Serve the right ad on eye contact They look. Camera profiles. Matching video plays — not a loop for everyone.

Brands pay for outdoor reach they cannot prove. With AdEye, an impression means someone actually looked — and once that eye contact is there, age, gender and style decide which creative from your pool goes on screen.

How it works

From glance to tailored ad

A camera built into the DOOH screen watches the footpath. Everything below happens on the on-board computer — in real time.

Camera Sees footpath and screen zone
Detect Find each passer-by
Profile Age, gender, clothing style
Attention Trigger when they look at screen
Play Best creative from your pool

From the concept demo

  1. A man in his 20s–30s in sportswear walks through a city square.
  2. He stops and looks at the screen. The camera detects sustained gaze.
  3. The system builds a brief audience profile and counts one qualified impression.
  4. The display switches to a Nike sneakers video matched to that profile.

Full privacy

All processing stays inside the screen

This is not a cloud service watching the street. The entire pipeline runs on a local computer built into the DOOH unit. That is what makes full privacy and anonymity possible.

Camera Sees the footpath
→
On-screen computer Detect · profile · choose ad
Real time, on the device
→
De-personalised stats Impression counts, dwell, segments

✕ No video to the cloud  ·  ✕ No images stored on device  ·  ✕ No face database

  • Nothing leaves the screen. No personal data transmitted to any cloud or external server.
  • Nothing is stored. Video frames are processed in real time and discarded. No recordings, no photos, no face data kept on the device.
  • Only anonymous statistics. De-personalised campaign metrics — e.g. one qualified impression, male, 20–30, sportswear, 2.4 s dwell. No identity.
  • Full anonymity. The system never knows who someone is. It sees a person in the moment, acts on broad attributes, and forgets.

Status

Proof of concept

This page presents the product concept. The technology behind it exists as a lab proof of concept — person detection, gaze recognition, audience profiling, and on-screen inference on standard billboard hardware. The next step is field testing on a live screen.