Video server

Real-Time surveillance face extractor.
The system captures and extracts face images from any video stream source: USB camera, IP camera, recordings, etc.

The video server is design to control 4 cameras on each server (2 cores per camera).

The video server is handling all the video streams from the cameras and consist of following components:

    •Face finder – find the faces in each video frame up to 4 faces per frame

    •Face cropping – isolate the face from the image and crop it

    •Tracker – track the face from frame to frame

    •Best shot analysis – decide the best quality face (from the sequence)

    •2D to 3D – decide if the face is frontal and if not rotate them accordingly

    •Tokenize – prepare the token image and send it to the application server


The application server is the system manager and consists of the following components:

IDT Middleware – a state of the art pre-processing component that correct the illumination and other image issues before the matching to enhance the results and the overall performance

Template creator – create the template from the given image

Watch list management – configure the system threshold and other system configuration settings

Notifications – provide the external system with the alarms for a positive match including the image from the camera and the image from the database, score, and other optional parameters

•The hardware of the application server can change and is depended on the number of cameras and calculation of the CPU that required for processing the above assignments

Matching engine

    •The matching sever holds the watch list and conducting 1:N searches

    •The matching results are send back to the application server

    •The application server will manage the results and reports them via the notification server to the external system

Mobile application

One of the most desired features of facial recognition systems is the ability to use the face recognition everywhere using a Smartphone application and the Smartphone camera.

In today’s technology the capabilities of the platform, the camera and the fast communication links give us the ability to use the facial recognition on the outdoor and on the move, we are now in the development phase of this capability that will allow the user to:

Take a picture of a subject face and:

    •Conduct database search to know if this subject is wanted or not
    •Enroll the subject to the system for future use.