Face Recognition Solution Based On i.MX RT106F

表示  限定  | 次の代わりに検索 

Face Recognition Solution Based On i.MX RT106F

Face Recognition Solution Based On i.MX RT106F

NXP MCU-level face recognbition solution is implemented by using i.MX RT106F, which makes the developers add face recognition capabilities to their MCU-based IoT products. This ultra-small size, integrated software algorithm and hardware solution can facilitate developers for rapid evalution and proof of concept development.

This solution minimizes time to market, reduces risk and reduces development work, which can make it easier for many OEMs to add face recogtion functions. It provides advanced user interface and access control functions for smart homes, smart appliances, smart toys and smart industries without the need for Wi-Fi and cloud connectivity, solving the privacy issues of many consumers.

i.MX RT106F is a member of the i.MX RT1060 series. It will be officially mass-produced in April 2020. It is mainly aimed at low-cost face recognition applications. It is based on the Arm Coretx-M7 core and a high-performance real-time processor with a frequency up to 600MHz.

In addition to the face recognition function, the i.MX RT106F also has a large number of available peripherals, which can be used as the main chip for a variety of applications. i.MX RT106F has been licensed to run NXP OASIS runtime for face recognition, including:

● Camera Driver

● Image capture and preprocessing

● Face Detection

● Face Tracking;

● Face Contrast;

● Face Recognition;

● Anti-fraud;

● Face Configuration;

● Confidence;

● Face recognition authenticat results;

● Built-in secure bootloader, application verification;

● Automatic Verification Script;

● Support MCUXpresso SDK, IDE and configuration tools.

Hardware Framework


Software Framework


Core Process of Software



Is customer able to modify the Oasis runtime library to suit their need if not, can they  implement their own library on RT106F? Can NXP library support human detection with other objects in the captured photo? If can, they can send the human photo to their server and process their own algorithm for better accuracy.

‎03-27-2020 02:16 AM