TECH5 launched its latest software solution based on T5-Face, one of the company’s core biometric technologies, with built-in facial mask detection, that will be used for touchless access control and time and attendance use cases.
Products based on T5-Face identify people in non-cooperative mode despite various potential artefacts such as faces partially covered with a mask or sunglasses, poor lighting conditions and others. In case of cooperative recognition, when the person is looking into the camera, and in some cases where recognition is running on mobile devices, the system may ask to pull down the mask such that a bigger portion of the face will be visible to facilitate fast and accurate verification or identification.
The solution allows for facial verification (1:1) as well as identification (1:N), detection of facial masks, dispatch of notifications to a central system for access management such as entrance permission and denial, among other functions. T5-Face SDK can be integrated with time and attendance systems as well as any mobile device with a built-in camera for access control and time and attendance.
“Due to the increasing demand for touchless biometric solutions during the pandemic, we are cooperating with our partners to provide reliable biometric-based systems tailored for the access control market. Our latest algorithm with built-in mask detection allows for facial recognition on mobile devices in a fast and accurate manner to make sure that access is granted only to those individuals who are entitled to” – says Rahul Parthe, TECH5 Co-founder and CTO.
TECH5 has already integrated the solution with devices of its hardware manufacturing partners, Wyse Biometric Systems, India and AndOpen, South Korea. Today, their systems are tested and successfully operating at access control points in a multitude of locations.
“When we started researching face algorithms to integrate with our product, we short-listed several technology providers, considering various criteria. TECH5’s products demonstrated successful enrollment performance and a very good verification accuracy on low-quality images, despite such artefacts as camera blur, bad lighting, and more. Also, it’s light enough to apply on low-performance CPUs such as ARM. Nevertheless, the performance of the algorithm is very satisfactory. The face recognition algorithm of TECH5 is very suitable for our new conceptual product which is performing double confirmation, liveness check, and ID card reading simultaneously” – comments Alfred Oh, Senior Vice President, AndOpen.
Currently, over 1,000 licenses have been successfully deployed and working for AndOpen, and more than 200 licenses have been deployed for Wyse with a high ramp-up expected in 2021 as touchless facial biometric becomes mainstream.
“After some initial tests, we did the first deployment of our system with T5-Face recognition software at a Premier Education Institution in Pune in June 2020. Encouraged by the response, we rolled out a further 70-80 T5-Face SDK-based recognition systems, partially replacing existing fingerprint-based systems as well as deploying new installations. Our deployments cover the entire spectrum of India, from Kolkata to Kerala. We also completed a successful trial with our devices in Kenya, overcoming our initial concern how the solution would perform on the African population. The system is working perfectly in all regions” – says Y. D. Wadaskar – Managing Director, WYSE Biometrics Systems.
Today, the WYSE attendance devices, integrated with T5-Face engine, successfully enrolled 20k+ faces across India. The devices perform 1:N matching at a given location against the local database. The company is also planning to introduce T5-Face engine enhancements to their existing visitor management system.