Our core technology adopts the the latest in bio-metric recognition. It uses DLA deep learning device to help speed up the recognition speed and accuracy performance. All facial analysis and identity recognition are completed in a tenth of a second, highly-efficient and simple.
It is SBI's premier facial-recognition system that can help filter and manage unknown subjects and dangers.
Facial detection through hardware, using Neural Net-work model to improve efficiency.
De-duplication of personnel, reducing the amount of external data transmission, lowering the burden on the server to greatly improve the efficiency.
Complete facial recognition in a fraction of a secondwith an energy-saving design to reduce power consumption specifications as a product advantage.
Specified system for real and natural recognition, canbe used on a single unit or integrated with other extended applications such as: access control, personnel management...etc to reflect the product's high-flexibility and practical function.
Pairing with a cloud system, with an easy-to-use browser interface one can easily control on the intranet or manage with a cloud system.
Personal information will not be stored on the unit, if the equipment is lost there's no need to worry about data being stolen or disclosed.
| Facial Recognition
✦ 100% zero-contact design
✦ 0.1s to complete recognition
✦ 99% high recognition, low false positive rate
✦ 2s to complete measurement
✦ ± 0.3°C high accuracy
The core of the product developed by SBI supports three main functions. Including, face detection and recognition, rapid temperature screening and cloud services.
Facial recognition is imported into Deep Learning Accelerator (DLA) for comparison and analysis. Highly accurate, no-contact temperature quick screening module for temperature measurement. And cloud service plays an important role in controlling back-end data and equipment with the advantages of fast speed with a low price point.
✔ Main Functions
• The overall system is based on FreeRTOS
• Facial recognition system: using DLA accelerator,
short recognition time (about 0.1 seconds)
• Temperature quick screen: adopt the no-contact
temperature quick screen module independently
developed by SBI
• Cloud service: It is independently developed by
SBI, which can quickly integrate Cloud
• Combined with IoT: can be used with SBI sensors
• Supports facial recognition, between 40 to 60 individuals.
• Supports 1:N facial authentication recognition, liveware
counterfeiting, face-mask identification
• Supports facial attribute assessment: including glasses, sunglasses,
blur and occlusion
• Supports server: can support larger enterprises
• Supports reduced redundancy function: fast detection, reducing server
and bandwidth burden
• Adopts 8*8 matrix thermal imaging sensor
• Temperature measurement area: forehead
• Temperature measurement speed: 2 seconds/per measurement
• Temperature measurement distance: effective distance 30 ~ 50 cm
• Temperature measurement accuracy: ±0.3°C (at 40 cm)
• Temperature measurement ambient temperature: 15°C ~ 30°C
• Sensor operating temperature: 15°C ~ 30°C
• Adopts Restful API/JSON format
• Supports system registration: add new servers quickly
• Adopts security system to safeguard personal information and privacy
• Cross-platform design: uses a standard browser to view data directly
• Background management: unifies control of equipment and personnel data in the background (attendance/ access control/ temperature)
• Supports CMOS Sensor: 640*480 resolution
• Supports button/pairing/LED display functions
• Supports speaker recognition (voice recognition)
• Supports connector: 4Pin IO connector
• Supports UART TX/RX 3.3V, Relay control output
• Supports Ethernet RJ45 connection/USB 5V power supply
• Mirrored-surface size: 7-inches, LCM: 4.3-inches
The product is used in front-end identification,
using Ethernet to communicate, no need to add
other hardware or equipment, simply use a browser
that can connect to the Internet to review information
and manage data in real time.
Website was started with Mobirise web templates