Go to documentation repository
...
Recommended scheme for theFacePSIM-based video surveillance system looks like this:
On this scheme Face detection is in use forface detection and vector generation on remote clients. Face Recognition Serverrecognition server is in use for online face recognition as well as face search in archive on the main client. The benefit of such division is that the more resource consuming process – vector generation – is performed on remote clients (their number can be increased). It allows creating a large system with common database of reference images and enlarging it at any moment.
All processes can be combined on the same computer for small systems (i.e. creation of Face detection and Face Recognition Serverrecognition server).
Moreover, there is a scheme when face detection is possible on remote clients only, and vector generation is performed on the main client (see item 7 of Configuring the captured faces recognition parameters section).
It is also possible to create a corresponding Face Recognition Serverrecognition server object for each Face detection on one computer, which is selected as an external capture for the Face detection (see Configuring Face Detection module). In this case, the detection and generation of face vectors using the Face detection objects will be performed in separate processes for each camera.
Note | ||
---|---|---|
| ||
Only one Face Recognition Serverrecognition server object can be created on one computer, which performs online recognition and face search in the archive. All other created Face Recognition Serverrecognition server objects can only be used to parallelize the detection calculations and the face vectors generation using the Face detection objects. |
...