Dahua Technology has unveiled its AI-driven distributed face recognition server, aimed mainly at city scenes or projects, such as safe cities, the security management of main traffic thoroughfares or plazas, business analysis for chain supermarkets, etc.
As a mass data platform for face management, the face recognition server is able to analyse image streams as structural data and store them in a distributed database. Its image comparison search engine technology also supports fuzzy search of large volumes of face data and provides results within seconds.
Supported by the integrated, high performing GPU P4 module and Dahua’s deep learning algorithms, the server can detect a human face with an accuracy of above 90%, along with detailed features such as age, gender, expression and glasses, which help identify subtle individual difference. It also supports multiple-person image search and orders by similarity according to the time and location information, providing a convenient way to retrieve information.
Comparisons with proactive alarms
After collecting facial data, the server also provides both dynamic and static image comparisons. Dynamic comparison supports a real-time 100-channel picture stream (about 8.5 million pictures) and 300 000 face control comparisons; while static comparison supports a 30 million picture comparison. Once a face that was previously defined is found in the enrolled database, it will send a notification or alarm immediately. The server stores a maximum of 300 000 human face pictures.
For more information, contact Dahua Technology South Africa, +27 72 628 6810, sales.za@dahuatech.com, www.dahuasecurity.com.
© Technews Publishing (Pty) Ltd. | All Rights Reserved.
printer friendly version