Skip to main content

Unique Person Counting

Measure true footfall by distinguishing new visitors from returning ones. Visylix Unique Person Counting combines person detection with Re-Identification embeddings to deliver de-duplicated visitor counts that reflect actual engagement rather than inflated headcounts.

About Unique Person Counting

Visylix Unique Person Counting solves the fundamental challenge of distinguishing repeat visitors from new individuals by combining person detection with Re-Identification embeddings. Unlike simple headcount models that inflate totals every time a person re-enters a camera's field of view, this model assigns a temporary anonymous identity to each detected individual and recognizes them across re-entries, multiple cameras, and extended time windows. The result is an accurate count of distinct individuals who visited a location during a given period, a metric critical for conversion-rate calculation, marketing attribution, and true footfall benchmarking. All processing is privacy-preserving, using only appearance embeddings without storing facial biometrics.

Key Features

Engineered for accuracy, speed, and seamless integration into enterprise workflows.

Re-ID De-Duplication

Recognize returning individuals across re-entries and multiple cameras using deep metric learning embeddings, ensuring each person is counted only once per time window.

Configurable Time Windows

Set de-duplication windows from minutes to full-day periods to match your business counting conventions, whether hourly unique visitors or daily totals.

Privacy-Preserving Design

Anonymous appearance embeddings replace facial biometrics, ensuring GDPR and CCPA compliance while still delivering accurate de-duplicated counts.

New vs. Returning Segmentation

Distinguish first-time visitors from repeat visitors and measure visit frequency, providing both unique-visitor counts and return-rate metrics.

Multi-Entrance Consolidation

Aggregate counts from all entry points into a single de-duplicated site total, preventing double-counting when visitors use different doors or gates.

Trend Reporting

Generate hourly, daily, weekly, and monthly unique-visitor trend reports with year-over-year comparison capabilities for strategic planning.

Real-World Use Cases

Deployed across industries to deliver measurable value from day one.

Retail Conversion Analytics

Measure true store footfall for accurate conversion-rate calculation against point-of-sale transaction data, eliminating inflated headcounts that distort business metrics.

Shopping Mall Tenant Reporting

Provide verified unique-visitor counts to tenants for lease negotiations, marketing spend allocation, and performance benchmarking across mall zones.

Museum & Attraction Analytics

Reconcile ticketed versus actual visitor numbers to detect undercounting, pass-sharing, or capacity leakage, and measure exhibit-level engagement accurately.

Corporate Campus Planning

Count daily unique occupants across campus buildings for space planning, cleaning schedules, energy management, and emergency preparedness decisions.

Technical Specifications

De-Duplication MethodAppearance-based Re-ID embeddings
Re-ID Accuracy (mAP)> 85% on Market-1501
De-Dup Time WindowConfigurable: minutes to 24 hours
Inference Speed< 20 ms per person (GPU)
Multi-Entrance SupportUnlimited entry points
Privacy ComplianceNo facial biometrics stored
Reporting IntervalsHourly / daily / weekly / monthly
Supported ProtocolsRTSP, RTMP, HTTP-FLV, WebRTC
HardwareNVIDIA T4 / L4 / A100 / Jetson
DeploymentOn-premise, edge, or cloud

Ready to Get Started?

Discover how Visylix AI models can transform your video infrastructure into an intelligent, automated system.