HUMAN LEVEL PERCEPTION

CORSIGHT

Autonomous AI

With leading artificial intelligence, Corsight platform enables accurate facial recognition, advanced analytics, anomaly detection.

Corsight utilizes Autonomous AI to unlock the power of this data by organizing it into clear, comprehensible intelligence – identifying concepts and patterns that yield actionable insight.

BEST FACE RECOGNITION TECHNOLOGY FOR SAFER CITIES

SMART CITY

TRANSPORTATION

AUTHENTICATION

STADIUMS & VENUES

RETAIL

INFRASTRUCTURE

FACE RECOGNITION PLATFORM

While typical state-of-the-art face recognition technologies are based upon facial landmarks or feature extraction, Corsight’s Autonomous AI creates dynamic signatures in a very high dimensional space that automatically identifies the unique patterns in each and every face.

Live Advanced Facial Recognition

Forensic Search for Investigation

High accuracy at large scale under problematic conditions

Highly scalable, multi-stream, simultaneously search by text, image, video or signature to instantly find the relevant footage

KEY FEATURES

Real-time detection for images and live video streams
Automatic N:N identification
Forensic analysis
Smart Search to find a specific clip or POI
Advanced analytics (e.g people counting, facial attributes)
Single image required for person enrollment
Works with any IP camera

SPECIFICATIONS

OVERVIEW

ROBUSTNESS	Highly accurate under problematic conditions including poor lighting, occlusions, low quality images and infrared light
POWERFUL AND LIGHTWEIGHT	The lightweight signature files preserve fast and high performance in large scale
ADDITIVE LEARNING	Fast adjustment to sub groups and conditions while not requiring full retraining
UNSUPERVISED LEARNING	Autonomously learn new faces online using single image only to create a database

PLATFORM

Software application
Multiplatform SDK for on-premise systems
Cloud API integration available

TECHNOLOGY

Signature Based Recognition
The Corsight system translates images into compressed signature files (lossless binary representations). No face images are stored in database. New images are translated to signatures and matched against the signature database for recognition. Zero images are saved or stored – allowing for enhanced compliance capacities.

MINIMUM H/W REQUIREMENTS