Click to order
Your order
Total: 
Your Name*
Your Email*
Company name
Tax number
Address
Payment method
* - Required field
 
© 2021 3DiVi Inc.
Case Studies

Papillon Access Control System

Access Control System with 3D Face Terminals

About Papillon

Papillon Systems is a vertically integrated provider of biometric solutions with annual sales of $20M. It supplies intellectual video surveillance and biometric access control systems to government and corporate customers across the world.

Website: papillonsystems.com

Problem

A coal mining holding in Kazakhstan wanted to upgrade its access control system with face recognition. Number of gates: 4, Number of employees: 1,200.

The system had to recognize faces "instantaneously" without interruption of flow during rush hours and had a built-in protection against spoofing by photo or video.

Solution

The Company served as an IT integrator in this project. The decision to utilize 3DiVi Face SDK was made because SDK includes and an advanced 3D liveness module.

The customer wanted 3D face recognition as it was known that 3D liveness works much faster and more reliably than 2D livensess solutions (despite the higher cost of 3D sensors, which adds about $100 for each terminal).

During the integration process, 3DiVi made recommendations on the selection of 3D sensors. After testing Intel RealSense and Orbbec Astra, Papillon selected RealSense for the Face Terminal: Intel NUC Mini PC was used as a processing unit. See close-up videos of assembled 3D Face Recognition Terminal:
To accelerate software development, pre-built components of 3DiVi Face Machine were used: Face Machine Server and Face Machine Client application. Face Machine Server (on-prem instance) stored face templates of employees, and Face Machine Client enabled face recognition on a terminal.

Our task was to develop a user interface for these components and to integrate it with the legacy HR management system of the customer.

During testing it showed excellent reliability (with zero false access errors on 1200 faces) and fast recognition speed (less than 300 ms per person). The project went live in 1Q2019 and development took about six months.
We would like to highlight an excellent error rate (zero so far on 1,200 database) and very fast recognition performance of the 3DiVi face recognition algorithm.
Georgy Bondarenko, Papillon Project Manager
Photo of installed Terminal

Products used

Face Machine: Face Machine Server (on-prem instance), Face Machine Client component running locally (with the 3D liveness module)