AI-driven Personnel Management and Access Control System June 7, - - PowerPoint PPT Presentation

AI-driven Personnel Management and Access Control System

June 7, 2019

1 2 3 Market Analysis Technical Solution Case Study 4 Market Scale

Market Requirement

Security

Provide higher security identity verifying method. Reduce the risk of card misuse, copy, lost, etc. Smooth verifying procedure. Better usage experience

Convenience Personnel Management

Flexible personnel permission management, such as employee, visitor, stranger, banned person, etc.

Labor Cost

Replace human verification by algorithm, decrease labor cost, and increase accuracy and efficiency

System’s Goal

Safer environment AI + personnel management Secure data management Smart building transformation

Transforming old buildings with advanced devices and infrastructures brings high technological elements to the user. Upgrading from manual data entry mode to automatic, secure, and sharable record storage, the user no longer has to worry about losing data. Smart devices coupled with the software help the user manage daily activities of personnel working/living inside in a new fashion. With advanced AI technology embedded, the platform helps to maintain a safe working/living environment.

1 2 3 Market Analysis Technical Solution Case Study 4 Market Scale

System Overview

Facial Recognition

• Face-recognition access control
• Largest face recognition platform in the world
• Financial security level
• Rapid, accuracy, non-cooperation capture mode

A B

Visitor Invitation/Guest-welcoming

• Mobile terminal for visitor register
• VIP-welcoming
• Visitor arrival reminder

C

Employee Attendance

• Facial recognition attendance
• Attendance statistic
• Attendance management

E

Megvii Koala Personnel Management Platform

D

Smart Surveillance

• Banned people recognition
• Stranger recognition
• Stranger alarm

System Customization

• Provide standard API to upper system
• Simplified development process

System Architecture

Platform Layer Application Layer Network Layer Access Layer

Megvii Koala Personnel Management Platform Access Control Visitor System Attendance System Guest Welcoming System Stranger Alarming Internet Intranet Face Recognition Turnstile Face Recognition Pad Face Recognition Camera Visitor Terminal

Network Topology

Intranet

Facial Recognition Server Facial Recognition Server Facial Recognition Server Management System Display Module Facial Recognition Turnstile Electromagnetic Gate Control IP Camera for Attendance, Welcoming, Stranger and Unwelcomed Guest Recognition

Application Scenarios

Government Building Office Building Sports Stadium Gym

Camera Height: 3m Recognition Distance: 12m Lenz: 50mm Camera Height: 5m Recognition Distance: 20m Lenz: 85mm

System Performance

Max database capability: 100,000 Recognition speed: <1s FMR: <0.5% Missing recognition in 3 frames: <1% Data-loading accuracy: >99% Quantity of facial processing per frame: 3-5 Support backlight processing

Access Control

Personnel management

• Centralized and unified information management
• Face image, name, department, occupation, and etc.

Access control equipment management

• Communication & operation status management
• Different switching strategies for different devices

Traffic records

• Display real-time results for all staff members
• Store all the entry, exit, and status records
• Record search available

Authority management

• Permissions for different personnel

Access Control

Electromagnetics Gate Turnstile M5 PAD M5 PAD

Visitor Management

Registration

• Visitor registration
• Appointment w/ records

Verification

• Self verification at self-service device
• Verify appointment & documents
• Grant access permission

Temporary Registration

• Visitors without appointment
• Manual registration
• Confirm with visited employee
• Verify documents
• Grant access permission

VIP Welcoming

• Customizable welcoming page

Visitor Management

Visitor Registration Visitor Recognition

Employee Attendance

Data collection and processing

• Facial recognition to take attendance
• Based on preset work shift
• Late arrival/early leave records

Parameter setting

• Custom national holiday setting

Interface

• Standard API interface
• Syncs with the user's HR system

Report statistics and query

• Monthly based
• Data export available

Smart Surveillance

Banned people recognition

• Custom watch list
• Turn on with a button
• Automatic alarm
• Large-screen display

Stranger recognition

• Automatic recognition
• Record storage

Quotation Model

Koala System H4 Server License C2/C3S-123 M5/M5E 1 1 10 10 50

• 2. H4/H5 Server
• 3. Quotation

Provided as per project requirement.

• 1. NUC Server

Koala System NUC Server License C2/C3S-123 M5/M5E 1 1 1 1 5

Hardware Spec-Smart Camera

Smart Camera Model: C3S-123 Pixel: 2 million，1/2.8” sensor Mainstream Resolution: 1920*1080, 1280*720, 704*576, 352*288 Sub-stream Resolution：D1（704×576） WDR:120DB

• Min. illumination: 0.0001 Lux(F1.2,AGC ON)

Shutter Speed: 1/100000s～1s Facial Recognition: Max. Facial image detection 50/frame； frame detection up to 25fps

Hardware Spec-C2 Camera

HD Camera Model: C2 Pixel: 2 million，1/2.8” sensor Mainstream Resolution: 25fps(1920*1080), 25fps(1280*720), Sub-stream Resolution：25fps(704*576), Video Compress:H.264

• Min. illumination: Color 0.001Lux@F1.2

Shutter Speed: 1/10s～1/1000s

Hardware Spec-NUC Server

NUC Sever Model: MegBox-B1D-122-1 Process Capability: Support access of 1 C2/C3S-123 camera or 5 M5/M5E PAD CPU: Dual core，1.9GHz Memory: DDR3 1600 8G(4G×2) GPU: Integrated GPU Harddisk: 128G SSD Interface: 1000M Ethernet card Dimension: 115×111×48(mm)

Hardware Spec-H4 Server

H4 Sever Model: KSS-DR1000-H4-3 Process Capability: Support access of 10 C2/C3S-123 camera or 50 M5/M5E PAD System: Ubuntu 16.0 CPU: Intel XEON E5-2630 V4*2 Memory: 8G DDR4*8 GPU: Leadtek Stock Edition Geforce GTX 1070 Harddisk: 1T *4 (RAID Card)+120GB SSD Interface: 10M/100M/1000M LAN port x 2 Max power consume：800W

Hardware Spec-M5/M5E PAD

M5/M5E PAD Model: MegID-W2K/MegID-W2K-E CPU: ARM-A17 4 cores GPU: ARM Mali-T764 RAM 2GB dual DDR3 ROM 16GB eEMMC OS: Android 5.1.1 10/100/1000Mbps ethernet，RJ45 interface Interface: mini USB, NO,NC,COM, Wiegand 26bit/34bit, POE+Power Supply Max power consume：20W Dimension: 248.5X130.5X28.1mm

Note: M5 is indoor model, M5E is outdoor model.

Hardware Spec-M5/M5E PAD

• Turnstile
• Vertical Installation
• Wall Mount

1 2 3 Market Analysis Technical Solution Case Study 4 Market Scale

CapitaLand

MEGVII provides an end-to-end solution of personnel management which based on private cloud platform. The system integrates tenant management system, visitor management system, and lift control system to provide comprehensive access management services for more than 20 commercial real estates of CapitaLand.

Aliyun Building

In 2015, Alibaba adopted the intelligent building and campus solutions of MEGVII, which enabled the park to achieve intelligence in the whole procedure from visitor invitation, visit registration, face recognition turnstile, floor guidance, and visitor welcoming. The company management, employees, and visitors can enjoy the benefits of intelligent office.

More Scenarios

Backlight Low Light Liveness Detection Quick Pass

More Cases

Office Building Retail Campus

1 2 3 Market Analysis Technical Solution Case Study 4 Market Scale

Market Scale – 2018 Sales data

Customer Type Quantity Commercial Properties 1338 Government Complex 662 Education 13 Offices 12 Customer Type Quantity Commercial Properties 1947 Education 730 Industrial Park 670 Offices 561 Governmental Complex 420 Retail/Business/Financing 128 Scenic spot 50

Direct Accounts Channel

Conclusion: Commercial properties is fundamental ，Education& Complex are increasing.

66% 33% 1% 0%

Commercial Properties Government Complex Education Offices

43% 16% 15% 13% 9% 3% 1%

Commercial Properties Education Industrial Park Offices Governmental Complex Retail/Business/Finan cing Scenic spot

Market Scale – 2019 Sales forecast

Competition

Commercial Properties 8000

Education 6500 Industrial Complex 4000 Governme ntal Complex 5500 Offices 2000

Others

Conclusion: Commercial Properties grow linearly, Education& Complex are blooming. Sales Amount

Power Human with AI