HID highlights image quality as key to effective facial recognition applications

HID is in the midst of a push with its facial recognition hardware and software for secure access control and retail payments, emphasizing the importance of image quality to successful implementations.

A blog post on high-performing facial recognition systems offers 10 pieces of advice for organizations putting the technology in place. Among them, HID highlights the importance of cameras with built-in “face-aware capabilities” and advanced capture technologies like multi-spectral imaging, which is used by HID’s U.ARE.U cameras. On the software side, algorithms must be trained with datasets that are demographically comprehensive, but also ethically sourced.

HID builds facial recognition algorithms from Paravision into its solutions, such as the Facepods integrated with Assa Abloy’s Speedgates at airports to deliver biometric flight boarding. The company has put together a list of questions for companies selecting a biometric system for air passenger processing.

High-performing cameras must also be tuned to work well in the real-world conditions they are expected to operate in, and integrate biometric presentation attack detection (PAD) capabilities. Data security measures beyond those offered by webcams are also needed to protect against data theft and injection attacks are also necessary, according to the post.

The post is focused on high-security deployments like ATMs and egates, but another post from HID explores the use of facial recognition in the self-service retail kiosks springing up in stores across the world.

Many of the same principles apply in enterprise access control and retail payments applications, but HID relates how its POS-maker partner Flytech builds retail tools into devices that integrate U.ARE.U cameras.

HID has also published a white paper on “Facial Recognition in Self-Service: Unlocking Its Full Potential” delving into the strategies retailers can use for unattended transactions.

Article Topics

biometric data quality  |  biometrics  |  facial recognition  |  HID  |  Paravision  |  U.are.U