The MediaTek Genio family of platforms has been trusted by device manufacturers around the world to deliver secure, robust, scalable and high-quality solutions for a wide range of applications. The platform of the MediaTek Genio series is a high-performance Edge AI IoT platform that can be applied to smart retail, industrial applications and smart homes. It also supports a variety of network connection methods to meet the needs of partners for a wide range of Internet such as Gigabit Ethernet, Wi-Fi 6 and 5G network modules.

The high-performance MediaTek Genio series platform can support multiple operating systems, including Yocto linux, Ubuntu OS, and Android, to meet the needs of partners in developing new products.

Star Vision Technology’s solution sharing this time mainly wants partners to understand the face recognition technology and integrate it into the MediaTek Genio series platform to achieve this. Facial recognition technology is developing rapidly in many fields. It is a technology that extracts algorithms and recognition models by analyzing facial features. Its basic operating principle includes capturing images, detecting faces, then extracting facial feature points, establishing an accurate face database, and comparing the quantified data with the face feature values ​​in the database to find the correct identity and present the results to be demonstrated.

This time, CyberLink FaceMe was specially used for demonstration on the MediaTek Genio series platform to achieve better demonstration effects. CyberLink has the world's top AI face recognition engine, which supports cross-platform operating systems (Windows, Linux, Android and iOS) and can run on most hardware devices. Whether it is high-end workstations, computers and mobile devices, or even IoT edge computing devices using low-power chips, FaceMe's accurate and reliable facial recognition technology can be quickly introduced into a variety of IoT and AIoT smart IoT devices and solutions, so FaceMe can be widely used in various occasions.

CyberLLink FaceMe has been optimized for MediaTek Genio's AI processor unit (APU). For test performance on Genio 350/700/1200, please refer to: Test performance of the FaceMe series.

Utilizing the precise recognition and high-speed computing performance of FaceMe face recognition, face recognition can be applied in many scenarios. Please refer toThe latest trends in face recognition applications, learn about 6 major industrial fields at once! [2023 latest version]

The application scenarios are as follows,

Smart factory:Automated attendance management and access control

Attendance system FaceMe TimeCLock: 

The clock-in interface is deployed on various devices with web browsers, and employees only need to swipe their faces to complete the clock-in.

Consumption Analysis FaceMe Smart Retail:

Analyze customer data, roughly where people come to the store are distributed, gender, age, mood and what kind of products they like, etc., to provide smart retail customer analysis solutions.

Next, I will demonstrate to my partnersFaceMeApplication

We can see that when first openingFaceMeAt that time, there were no new people added to the database, so there was no way to recognize faces in the screen.

Just click on the screen to start adding people
 s

After the addition is completed, you can instantly check on the screen whether the person has been recognized.

After adding the personnel list, you can directly apply it to the database through database transfer.MediaTek GenioImplemented on each platform of the series, administrators no longer need to rebuild the database.

Take the database built on the Genio700 platform as an example and migrate it to the Genio1200 platform for use.

Click Database Management in settings

Select the folder to export the database to [EXPORT]

Next, place the created database inGenius 1200Within the platform, through the ADB command

C:platform-tools>adb root
restarting adbd as root
C:platform-tools>adb shell
aiot8390p6_64_bsp:/ # cd /sdcard/
aiot8390p6_64_bsp:/sdcard # ls
Alarms Audiobooks Documents Movies Notifications Podcasts Ringtones test_pic.jpg
Android DCIM Download Music Pictures Recordings test.mp4
aiot8390p6_64_bsp:/sdcard # cd DCIM/
aiot8390p6_64_bsp:/sdcard/DCIM # ls
faceMe_out.fdb
aiot8390p6_64_bsp:/sdcard/DCIM #aiot8390p6_64_bsp:/sdcard/DCIM # pwd
/sdcard/DCIM
aiot8390p6_64_bsp:/sdcard/DCIM # exit


C:platform-tools>adb pull /sdcard/DCIM/faceMe_out.fdb .
/sdcard/DCIM/faceMe_out.fdb: 1 file pulled, 0 skipped. 8.0 MB/s (114688 bytes in 0.014s)

C:platform-tools>
C:platform-tools>adb push faceMe_out.fdb /sdcard/DCIM
faceMe_out.fdb: 1 file pushed, 0 skipped. 30.2 MB/s (114688 bytes in 0.004s)
C:platform-tools>
C:platform-tools>adb shell
aiot8395p1_64_bsp:/ $ su
aiot8395p1_64_bsp:/ # ls /sdcard/DCIM/
faceMe_out.fdb
aiot8395p1_64_bsp:/ #

[IMPORT]

After loading the database, partners canGenius 1200Previously added personnel are directly identified on the platform

本That’s it for this drill. Friends who are interested can study and discuss together. Thank you.!

►Scenario application diagram

►Solution block diagram