Vision Analytics

Proof of Attention,
Not Guesswork

Adflux Vision Analytics delivers privacy-first audience measurement for your retail media network. Choose the solution that fits your infrastructure — with or without cameras.

Retail media vision analytics uses AI computer vision cameras or Bluetooth sensors to measure audience impressions, dwell time, and demographic data in front of in-store digital advertising screens. This data enables retailers to sell verified, audience-based advertising to brands — replacing estimated impressions with actual measured exposure.

AI Vision Analytics

Camera-Based Audience Intelligence

Adflux Vision Analytics with cameras delivers the most advanced audience measurement available for retail media networks. Using edge-processed computer vision, the platform anonymously detects and analyses audience behaviour — converting physical presence into actionable advertising intelligence without capturing or storing any personal data.

Integrated directly with the Adflux CMS, the system enables content effectiveness measurement, dynamic content triggering, and performance-based pricing models — giving advertisers proof of attention, not just proof of play.

Edge ProcessingNo PII CapturedCMS IntegratedReal-Time Triggers
Vision Analytics
1,100
Impressions
856
Views
657
View-Throughs

Anonymous Audience Measurement in Action

Edge-processed computer vision anonymously detects and classifies audience members — no personal data is ever captured or stored.

Anonymous Audience Measurement

Analytics Maturity Curve

Three progressive stages of audience intelligence — from measurement to automation.

01Descriptive

Anonymous Audience Measurement

Capture anonymous audience data — demographics, age group, gender, dwell time, and attention metrics — without storing any personal information.

02Descriptive

Content Effectiveness

Combine audience data with proof-of-play data from the Adflux CMS to deliver eyeball-to-content measurement. Understand which creatives drive attention and engagement.

03Predictive / Proactive

Content at the Right Opportunity

Set rules to automatically trigger content based on audience attributes or time of day. Deliver the right message to the right person at the right moment.

Content Effectiveness Measurement

Combine audience presence data with proof-of-play from the Adflux CMS to deliver true eyeball-to-content measurement — understanding which creatives drive real attention and engagement.

Content Effectiveness

Metrics Delivered

Every metric is anonymised, edge-processed, and integrated with your Adflux CMS campaign data.

Impressions
Total verified audience exposures per screen and campaign
Views & Glances
Differentiated attention events — full views vs brief glances
Dwell Time
Average time audience members spend in front of a screen
Attention Time
Duration of active engagement with displayed content
Content Effectiveness
Percentage of audience who engaged meaningfully with content
Top Persona
Dominant audience demographic for each screen and zone

Edge-to-Cloud Data Pipeline

Raw video is processed entirely at the edge device. Only anonymised metadata tokens are transmitted to the cloud — raw video never leaves the premises.

Edge to Cloud Data Flow

Edge-to-Cloud Data Flow

Raw video is processed at the edge — only anonymised metadata ever reaches the cloud.

Camera Sensors
Capture objects and environmental data
Edge Processing
AI models extract anonymised metadata on-device
Secure Transmission
Only anonymised tokens sent to cloud (HTTPS)
Cloud Analytics
Aggregation, transformation, and advanced analytics
Insights Delivery
Visualised reports integrated with Adflux CMS

Technical Architecture

The platform supports three deployment tiers — Lite, Standard, and Enterprise — each with a corresponding infrastructure footprint from edge device through to cloud analytics.

Technical Architecture

Camera Hardware Options

Adflux Vision Analytics supports two certified USB RGB sensor options, each paired with a compatible edge device.

USB RGB Sensor — Option A

Image Sensor1/2.7-inch Full HD CMOS
Field of View95.8° Diagonal
Resolution1920×1080 at 60fps
Notes90cm focus distance

USB RGB Sensor — Option B

Image Sensor1/2.7 CMOS, High Dynamic Range
Field of View144.2°(D) × 121.2°(H) × 64.2°(V)
Resolution1920×1080 at 60fps
NotesSuperior low-light performance
Edge Device
Intel Core i5-1235U · 8GB RAM · 128GB+ Storage · Windows 10 IoT LTSC
Camera Placement
Optimal height: 1.6–2.0m · Detection range: 1.6–10m · Tilt: 11–15° downward
Network
Wired Cat5/Cat6 preferred · 1 camera per edge device · Internet connection required

Privacy-First by Design

Raw video never leaves the device. All analytics are anonymous, aggregated, and non-identifiable.

No facial recognition — ever
No personal data captured or stored
Edge-processed: raw video never leaves the device
Anonymous Person-Like-Object (PLO) detection only
No racial bias — AI trained entirely on synthetic data
UNESCO Ethical AI guidelines compliant
TRUSTe Privacy certified
GDPR, ISO 27001, APEC Privacy Framework aligned
Australian Privacy Act 1988 compliant
Get in Touch

Request a Vision Analytics Demo

Tell us about your network and we'll show you the right solution.