Go to documentation repository
When you change the used face recognition module, for example, replace the Cognitec module with the Tevian the VA module or vice versa, the biometric vectors are automatically re-generatedregenerated. This is necessary because biometric vectors received by one face recognition module cannot be used by another.
The biometric vectors are also automatically re-generated regenerated when you switch the Tevian VA SDK version or the face recognition and search algorithm of Tevian and of VA and VideoIntellect 1.1 face face recognition modules.
| Info | ||
|---|---|---|
| ||
To switch between the used face recognition modules, use the Tevianthe Va.IdentificationAlgorithm registry key (see Registry keys reference guide, for details about working with the registry, see Working with Windows OS registry). |
The re-generation regeneration of biometric vectors is performed in the background without disturbing the face recognition system operation. Firstly, the faces which are already entered in the database are re-generatedregenerated. The current process of biometric vectors re-generation regeneration is displayed in the Face recognition and search interface module, on the Statistics tab, in the Faces in RAM field (see Displaying the statistics). If the Server is restarted, the process of re-generating regenerating the remaining faces will continue automatically.
| Info | ||
|---|---|---|
| ||
By default, the biometric vectors are re-generated regenerated on one CPU thread. Using the Update.CpuCoreCount registry key, you can set an arbitrary number of threads on which biometric vectors will be re-generated regenerated (see Registry keys reference guide). |
You can also re-generate regenerate the biometric vectors manually. To do this, do the following:
Wait for completion of biometric vectors re-generation regeneration.
| Note | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |||||||||||||||
Errors may occur during the biometric vectors re-generation regeneration process. For example, the an error below may occur due to the fact that the new face recognition module SDK or the algorithm of the Tevian VA ALG2 face recognition module cannot recognize a face on a poor quality image and recreate a vector due to its better and more demanding recognition capabilities. Therefore, such image can be skipped.
If Axxon PSIM malfunctions after the vectors re-generation regeneration, do the following:
|