NXP has unveiled a new low-cost facial recognition solution that will make the technology more affordable for equipment manufacturers. The solution is based on NXP’s i.MX RT106F crossover microcontroller (MCU), and comes with a license for the AI and Machine Learning software needed to run facial recognition.
The NXP offering is compatible with infrared and RGB cameras, and with flash and SDRAM memory systems. It offers liveness detection for additional security, and delivers everything at a total cost lower than $10 per unit. The MCU itself goes for $4 per unit at a volume of 10,000 units, while the SLN-VIZNAS-IOT secure face recognition development kit is available for $199.
According to NXP, the dev kit dramatically reduces the time to market for device manufacturers, allowing them to finish their designs in as little as six months. To that end, the kit comes with schematics, a board layout, and a bill of materials (BoM) in addition to the facial recognition algorithm itself.
The MCU can be deployed for digital and physical access control across a variety of channels. For example, it can be used to limit access to certain machines or dangerous equipment, or it can be integrated into a security system that covers doors, elevators, and other points of entry.
The facial recognition scan takes place on the edge rather than in the cloud, with a process that takes less than half a second. The fact that biometric information does not get sent to the cloud also helps to ensure data privacy. Biometric information can be registered remotely using a PC or mobile application.
The i.MX RT106F debuted in September of 2019, and is designed to support offline facial recognition in smart devices. NXP followed that with a similar MCU for voice recognition, and has since continued to release new MCUs built for IoT and smart home applications.
November 6, 2020 – by Eric Weiss