Face-RK3399

Face Recognition Main Board

Face-RK3399 is small-scale and with strong face recognition performance,supports multiple face recognition algorithms and various recognition modes. Its rich peripheral interfaces can carry all kinds of peripheral equipment of gate, and meets the application needs of access control, attendance and paying with faces. It also can provide one-stop solution service to help projects land quickly.

“Server-level” Strong Processor

Uses Rockchip six-core 64-bit “server-level” processor, frequency up to 1.8 GHz, integrates quad-core Mali-T860 GPU. It supports multi-format video coding and decoding, and has better processing performance.

Efficient Identification And High Accuracy

Supports various face recognition algorithms smoothly, and the response to detection speed is efficient. Recognition speed is less than 300 ms and recognition accuracy is more than 99.77%.

Support Multiple Recognition Modes

Maximum supports identify 100,000 face database and multiple recognition modes(1:1, 1: N, M: N) to realize multi-person recognition and detection at the same time easily.

Carry Mainstream Equipment

Face-RK3399 can carry a variety of peripheral safety equipment of gate, as well as supports mainstream cameras in the industry (binocular, structured light, monocular), MIPI interface display screen, ID card reader, QR code scanner, IC card reader, fingerprint reader, and other equipment.

Small-scale And High-performance

Face-RK3399 with beautiful and compact layout, the size is only 103mm x 64mm. It saves more valuable space to be embedded in various devices easily.

Form Face Recognition Machine

Equipped with mainstream cameras of industry (binocular, structured light, monocular), enclosure, display screen and so on, which can form a high-performance face recognition machine quickly.

Equipped with Management System

Equipped with face recognition program (APK), PC background management software and mobile phone management software (WeChat APP). They can be directly applied to the project by matching the whole machine.

Open Source

Open supporting Android SDK source code, face recognition API, development tutorials, development tools, and other materials to meet user’s self- exploitation needs.

Rich Extension Interfaces

Face-RK3399 supports Wiegand interface (26bit)/RS485 (switching through relay), LED driver interface (white light, infrared light), I2C interface (can be connected to TP and light sensor meanwhile), supports USB/MIPI interface of monocular, binocular and structured light camera and 4G LTE network. It can carry various external devices easily.

Applications

Face-RK3399 can be widely used in buildings, communities, campuses, construction sites, factories, etc.

  • Office building

  • Communities

  • School

  • Factories

  • Subway

  • Store

Specification

unit:mm
Specification
SoC

Rockchip RK3399 (28 nm HKMG manufacture procedure)

CPU

Six-core 64-bit processor (dual-core Cortex-A72+quad-core Cortex-A53), frequency up to 1.8GHz

GPU

Mali-T860 quad-core GPU, Support OpenGL ES1.1/2.0/3.0/3.1, OpenVG1.1, OpenCL, DX11

VPU

Support 4K and 1080P H.265/H.264 video decoding, up to 60fps

RAM 2GB LPDDR4(2GB/4GB)
Storage

16GB eMMC 5.1 (16GB/32GB/128GB), Support TF card expansion

Hardware Features
Ethernet

10/100M Ethernet interface

WiFi

2.4GHZ single frequency WiFi, support 802.11 b/g/n

Display

1 x MIPI-DSI, Single channel supports up to 1080P@60fps output (support 5"/7"/8"/9.7"/ 10.1" display)

Audio

1 x Speaker audio output (4Ω 3W)

1 x MIC audio input

Camera

Support monocular camera, binocular camera, Structured light camera

MIPI camera: Supports 2-way MIPI camera (built-in dual hardware ISP, up to single 13Mpixel or dual 8Mpixel)

USB camera:Supports 2-way USB 2.0 camera

USB

USB 2.0 × 4

(USB 2.0 Wafer socket × 2, 20Pin: USB Host × 1, USB OTG × 1)

Fill Light

3 x LED drive interface (white fill light), infrared fill light

Socket

Light sensor × 1, touch screen × 1, RTC × 1, ADC × 1, Debug × 1,

5V power output × 1, 12V power input × 1

Key

Reset×1, Power×1, Recovery×1

External Interface(20Pin )

1) DC 12V input

2) USB HOST

3) USB OTG

4) 10/100M Ethernet

5) RS485, Wiegand interface (switched by relay)

Power

DC 12V - 2A

Software and Support
OS

Android 、Ubuntu

One-stop Solution

(1)Provide complete face recognition program, background and mobile management program

(2)Provide USB/MIPI monocular, binocular, structured light camera (optional)

(3)Provide 8 inch/ 10.1 inch IPS screen (optional, support capacitive touch)

(4)Provide aluminum alloy case (based on 8-inch display, binocular camera)

Open Source

Android SDK source code, face recognition API

Features
Recognition Accuracy

>99.77%

Face Detection

Support up to 32 simultaneous detection (monocular camera)

Recognition Distance

1.5m

Face Anti-Spoofing

Support

Recognition Mode

Support 1:1, 1:N, M:N multiple modes

Face Database

Support up to 100,000

Appearance
Size

103 mm×64 mm

Operating Temperature

0℃-60℃

PCB

6-layer board design