Falcon

Human Recognition System

Features

  • High-performance, High-capacity
    • Supports up to 500,000 cameras in one system
    • Supports up to 10,000,000 Face / Human Detection Comparison Library
    • Real-time multiple face matching, matching speed low to 2s
    • Up to 98% face recognition accuracy
  • Deep Analysis
    • Supports deep learning algorithm
    • Supports human attributes analysis
  • Good Compatibility
    • Supports different types of camera: Kedacom human recognitive camera, Kedacom standard IP camera, Third-party IP camera
  • Abundant Functions and Features
    • Three types of snapshot: face, head and shoulder, person
    • Supports recognizing and capturing the human side and back picture
    • Supports searching picture by picture or attributes

Technical Specification

Part Number

  Falcon Human Recognition System

KProServer3 CPU Server

CPU Server Chassis High density server chassis
19″ 2U standard blade rack server chassis, able to insert 4 computing nodes, each computing node configuration 6 slot of 3.5 inch / 2.5 inch compatible hard drive slot, a total of 24 hard disk slot can be placed
Include 2 unit of 2150W redundant platinum power supply
Computing Nodes Computing Nodes for KProServer3-B01
2 x Intel Xeon Silver 4110 (8 core, 16 thread, up to 3.0GHz, 11MB Cache),
32GB RAM (KProServer3-N100-E01-32G)  / 64GB RAM KProServer3-N100-E01-64G) / 128GB RAM (KProServer3-N100-E02-128G)
1 x 2.4TB SAS, 1x 960GB SSD, 2 x SFP
OS license not included
Functions Compares the attributes of each face
Calculates the similarity percentage between faces
Operating Temperature 10°C ~ 40°C / 50°F ~ 104°F
Operating Humidity 50% ~ 90%
Power 100 ~ 127V AC, 50 ~ 60Hz / 200 ~ 240V AC, 50 ~ 60Hz
Power Consumption Max. 2150W (4 computing nodes included)
Weight 38kg / 83.78lb (including 4 x nodes)
Chassis Dimensions 733 x 438 x 86.9mm / 28.86’’ x 17.24’’ x 3.42’’

DeepEngine-1000 GPU Server

GPU Server Chassis High density server chassis
19″ 4U standard blade rack server chassis, able to insert 6 computing nodes
Include 4 unit of 1200W redundant platinum power supply
Computing Nodes Computing Nodes for DeepEngine-1000
1 x 64 bit multi-core CPU, 2 x GPU, 32GB RAM
2 x 1000M RJ45, 240GB SSD
OS license not included
Functions Identifies the faces and persons in the video streaming
Grabs the attributes of each face and person
Operating Temperature 0°C ~ 40°C / 32°F ~ 104°F
Operating Humidity 50% ~ 90%
Power 100 ~ 127V AC, 50 ~ 60Hz / 200 ~ 240V AC, 50 ~ 60Hz
Power Consumption Max. 3000W (6 computing nodes included)
Weight 58kg / 127.87lb (including 6 x nodes)
Chassis Dimensions 820 x 440 x 177.8mm / 32.28’’ x 17.32’’ x 7’’

System

Supported Operating System Windows10 Professional 64-bit / Windows 10 Enterprise 64-bit
Windows 7 Professional 64-bit / Windows 7 Enterprise 64-bit / Windows 7 Ultimate 64-bit
Windows Server 2008 / Windows Server 2012
Database Elastic Search
Application System Falcon Human Recognition System
User Management Admin / User
Maximum Number Of Camera / Video Stream Access Up to 500,000
Face / Human Detection Comparison Library Up to 10,000,000
Camera Access Type Kedacom human recognitive camera
Kedacom standard IPC (by Kedacom VMS)
Third-party IPC (by Kedacom VMS and NVR)
RTSP Video Stream (Optional)
Storage Face and human pictures (By IPSAN)

Recognitive Functions

Target Detection Identifies the faces and persons in the video streaming
Captured Picture Person, Face, Full view
Face Matching Real-time / Post-event
Analysis Analyzes attributes of human and face image, grabs the gender, age group, with or without glasses and mask, color of clothes, texture of clothes, direction, speed level
Search Mode Date and time / Camera / Face image / Person image / Attributes (gender, age group, with or without glasses and mask, color of clothes, texture of clothes, direction, speed level)
People Counting Bi-directional

Alarm

Alarm Triggers Face matching
Alarm Events Alarm report / Snapshot / Acoustic alarm

Video

Compression H.265 / H.264
Live View 1080p@30fps

Network

Network Protocols TCP/IP, UDP, RTSP, VSIP, 802.1x, IPv4
User Interface EPAS Client

Resource

Specification Document

Falcon Human Recognition System Specification Document