AI visual algorithm module, the main chip uses AI main chip, quipped with two 2-megapixel HD cameras.The AI visual algorithm module uses Linux Quick-start operating system,enable fast startup. The AI visual algorithm module has locally embedded binocular 3D liveness face detection algorithm, deep neural network recognition algorithm based on big data training, QR code recognition algorithm, palm print and palm vein algorithm, realizing face, QR code, palm print and palm vein recognition functions, and outputting recognition results through UART. At the same time, USB can transmit color camera UVC video and UAC audio for local audio and video display and playback or remote video and audio transmission functions.
The AI visual algorithm module has a fast startup speed and supports both low power consumption and constant power supply modes. In the normal power supply mode, when the module is set to the low-power automatic detection mode, it can realize the automatic detection and identification of face, palmprint vein and two-dimensional code and actively report the function, which is convenient for the docking of third-party products. It can enable third-party products to add multiple identity authentication functions such as face, palm print, palm vein, and QR code.
The AI vision algorithm module provides UART protocol, supports third-party integration and docking, and realizes audio and video transmission docking of identity recognition function.
The AI vision algorithm module has no computing power requirements for third-party devices, is plug-and-play, and has strong versatility. It can upgrade new and old products to meet various identity authentication needs such as face, QR code, and palm print.All kinds of new and old products that need face, two-dimensional code, palm print and other identity authentication can be upgraded.
ØSupport binocular 3D liveness detection to prevent fake face attacks such as photos and videos;
ØSupport neural network deep learning face recognition algorithm;
ØSupport palm print and palm vein algorithm
ØSupport QR code algorithm;
ØStandard face users 2500,+palmprint users 2500. Can be customized single face users 5000 ;
ØSupport 1:1 and 1:N recognition;
ØSupport face photo sending and registration;
ØSupport feature code upload and sending functions;
ØSupport recognition in complete darkness and outdoor non-sunlight;
ØSupport UART intercommunication functions for main functions such as recognition results, records, and registration (see the agreement for details)
ØNo computing power requirements for third-party devices, no need to add algorithm SDK, AI vision algorithm module local offline recognition,
ØMatch professional customized dual cameras, customers do not need to add cameras
ØSupport continuous iteration and upgrade of algorithms
ØSupport customization of other AI vision functions
|
Product name |
AI vision algorithm module |
|
Product model |
AI-50 |
|
SOC |
CPU: Arm 900M, NPU: 0.5T, RISCV: 600M; |
|
DDR2 |
128MB |
|
FLASH |
32MB |
|
Image Sensor |
2, 1/5” COMS, 2 million pixels, dual MIPI interface |
|
LENS |
FOV: Diagonal 83°, ideal focal length: 60CM |
|
LED |
IR: 850 wavelength, RBG: 650 wavelength, 90°. |
|
USB interface |
Support USB video transmission (UVC) function Extensible support audio and video (UAC) function, support U disk upgrade function, support MJPG (extensible support H.264, YUY2) encoded image output. |
|
UART |
External communication interface |
|
Communication baud rate (UART) |
115200 (default) |
|
Power supply voltage |
5V~12V(>=1A) |
|
Operating current |
320-330mA@8V |
|
Standby current |
120-130MA (in automatic detection mode) |
|
Shutdown current |
0uA (in non-automatic detection mode, power off when not in use) |
|
Module startup time |
900ms~2.5s(The startup speed is related to the user's storage capacity |
|
Operating temperature |
-20°C ~ 60°C |
|
Storage temperature |
-30°C ~ 70°C |
|
Relative humidity |
10% ~ 93% (No condensation) |
|
Speaker/mic |
Extensible support |
|
Number of users (face&Palm) |
Default:2500palms+2500faces), customizable:5000faces |
|
Supported algorithm |
Support binocular 3D liveness detection algorithm (standard); |
|
Face liveness detection false acceptance rate (LDAFAR) |
≤1% |
|
Face liveness detection false rejection rate (LPFRR) |
≤1% |
|
Pass (recognition) rate (face) |
98.85% |
|
False recognition rate (face) |
0.001% |
|
Recognition angle (face) |
About 20 degrees left and right, up and down (supports multi-angle face entry to expand the recognition range) |
|
Recognition distance (face) |
30~120cm |
|
Recognition distance(Palm) |
optimum distance :15CM |
|
Recognition distance(QR code) |
optimum distance :15CM |
+86-0755-2607 5207
