Face Facial Recognition Development System Based on Intell® VAS Algorithms

AIoT

WPI Intel IoT Solution Aggregator Team

Description

(Please download to learn more about this test briefing)

Face identification is a form of biometric authentication that relies on detection of facial features to verify a person’s identity. Technologies that involve automatic detection, tracking and identification of photographed images or video recordings containing people’s faces captured on cameras or video recorders all fall under the category of face identification or face recognition. Such technologies are widely used in finance, law enforcement, military, public safety, border control, government, aerospace, electrical power system, factory management, education, healthcare and a slew of business applications. With face recognition technologies continuing to advance toward maturity while gaining social acceptance, their applications will keep expanding to the following areas, just to name a few.

Security and management of commercial and residential premises, for instance, face recognition access control and time attendance systems or face recognition door locks.
E-passports and e-ID cards. First Research Institute of the Ministry of Public Security of PRC is expediting the planning and implementation of biometric passports.
Public safety, law enforcement and crime investigation, for example, searching for suspects using face recognition systems and nationwide networks
Self-service such as facial recognition payment systems

(Please download to learn more about this test briefing)

Resource Application

Solution overview

Intel® VAS algorithms

Developed by Intel’s algorithm department, the Intel® VAS (Video AnalyticsSuite) algorithm enables the development of high-precision and high-reliability face recognition, facial feature analysis, face comparison and detection and face tracking systems based on AI deep learning technologies in Windows 10 and Ubuntu 16.04 environments. Upon the signing of a non-disclosure agreement, customers can use the Intel® VAS algorithm at no cost to quickly develop their face recognition applications and products. If the products are based on Intel® CPU, customers can add smart features to their products to give them a competitive edge without increasing hardware costs.

JWIPC E7QL industrial computer

Featuring premium specs and long‐term operational stability, JWIPC’s E7 series is designed specifically for industrial application and is widely used to carry out many types of visual inspection work.

VAS installation guide
First, install Ubuntu 16.04 and Intel® OpenVINO and set OpenVINO environment variables.
1. Under Linux, install VAS using the file “vas_2018r3.789086a93859567_ubuntu_16.04.zip.” Unzip the file to the development directory, for example, "~/work/Intel®_vas_package/work/" as illustrated below.

2. Use an editor of your choice to add a line (as shown below) to “bin/setvars.sh.”
export PATH_VAS="$VAS_ROOT"