NXP Semiconductors has announced the world's first microcontroller-based offline face and expression recognition solution designed to provide facial and expression recognition processing for smart home, commercial and industrial devices. Based on NXP's latest cross-border MCU series i.MXRT106F, this face recognition solution helps customers quickly, easily and cost-effectively integrate facial, expression and emotional recognition functions into various IoT applications by running the FreeRTOS operating system.

Based on NXP's OASIS face processing engine, the MXRT106F has breakthroughs in both price and performance, using neural network algorithms for face detection, identification and protection against fraud, without the need to connect to the cloud. Customers can now use NXP's excellent hardware and software platform to provide advanced human machine interface (HMI) capabilities for end users' personalized experiences in smart appliances, thermostats, lighting systems, alarms, power tools and many other smart edge devices.

The MCU-based face recognition solution delivers the precision and low latency required for face and expression recognition, combining everything in an ultra-small form factor for a wide range of existing applications. The entire platform includes mass-produced pre-certified hardware devices and software, as well as the OASIS face processing engine fully integrated with the RT106F. The lens and display driver are used to identify the user's face and expression. This total solution reduces developers' barriers to expertise, supply chain and logistics management.

Compared to application processor solutions, RT106F-based solutions can save half or more of total system cost. The solution will be fully available in the first quarter of 2020, and NXP is currently working with OEMs to launch an evaluation and development kit for the solution as soon as possible.