Technical Capabilities

FLEXIBLE SYSTEM FEATURES
  • Full system configuration UI for admin and users

  • Configurable alarms, alerts and messaging

  • Fast checking and alerts. Subsecond alert generation; 2s/12m record search

  • System threshold and false positive management via admin portal or SDK VMS integration (including deeply integrated Milestone plug-ins)

  • Access control integration (eg: Gallagher, Integriti, Inception…)

  • Database management using PostgreSQL or SQLite

  • Bulk enrolment options for database integrations

  • Extensive reporting capabilities

  • Grayscale and colour support

  • Formats supported include but not limited to JPEG, PNG, WebP, H264, H265, MJPG, MPEG4

  • Operates in stand alone or integrated modes

 

 

DEPLOYMENT OPTIONS
  • On premises

  • Distributed

  • Headless (services)

  • SDK

  • Web portal

  • Cloud


EXTREMELY ROBUST AGAINST
  • Expression changes

  • Large pose changes

  • (+/- 90° yaw, +15/-30° pitch, +/-30° roll)

  • Beards, moustaches and hairstyle changes

  • Facial obscuration

  • Glasses, hats and headgear

  • Lighting variations

  • Low/high-quality images

  • Indoor/outdoor operation

  • Motion blur

  • Compression artefacts

CORE PRODUCT CAPABILITIES
  • Facial Detection

  • Facial Recognition

  • Face Verification (1:1)

  • Face Identification (1:many)

  • Mask detection / recognition

  • Facial onboarding / enrolment

  • Facial capture by video stream, image upload/scan, and web capture

  • Face analytics and reporting – age, gender, emotion, mask detection by time/date/location

  • Video analytics – crowd counting, people counting, tracking by time/date/location

  • Emotions (at present): Angry, Sad, Fear, Happy, Surprised, Disgust, Neutral

  • Gender 95% accuracy

  • Age groups: - currently 0-13, 13-20, 20-35, 35-45, 45-60 and 60 +

  • Known/unknown associate identification

  • Forensic analysis and reporting

  • Anti-spoofing options

  • Facial recognition based on advanced deep learning algorithms

FEATURE EXTRACTION
  • Ability to generate tiny face identification models representing a person’s unique biometric

  • Ability to extract information while not keeping a record of the image if required

  • Accurate video recognition using multiple frames

  • Combines multiple face images of any given person to generate

FACE DETECTION AND TRACKING
  • GPU-accelerated real-time face detection

  • Users can define regions of interest for face detection

  • Ability to determine landmarks in less than 1 ms for extremely accurate face alignment

  • Ability to track various faces through multiple types of video streams such as webcam, CCTV, archived video images etc.

  • Generate a unique tracking ID for every person and delivers face locations in the video stream