FullAnalysisRate
Specifies how frequently to recognize new objects, so that you can optimize performance for your hardware.
When you ingest video, object class recognition has to recognize new objects and track objects through the scene. After objects have been recognized they are tracked across every sampled video frame, but to improve performance Media Server only attempts to recognize new objects every N sampled frames (where N is specified by this parameter).
For example, if you set FullAnalysisRate=4
, Media Server recognizes all of the objects present in the first sampled frame. Media Server then tracks those objects across the next three frames. Media Server does not attempt to recognize new objects until the fifth sampled frame.
The amount of time between sampled video frames is controlled by the SampleInterval configuration parameter. If you need to choose between increasing this parameter and increasing the sample interval, Micro Focus recommends increasing this parameter first.
The default value of this parameter is greater for second-, third-, and fourth-generation recognizers because those recognizers usually perform better. (Object class recognition is faster, especially if you use a GPU, and the chosen sample interval is usually smaller). In other words, second-, third-, and fourth-generation recognizers should provide better performance, even when recognizing new objects on a smaller proportion of the analyzed video frames.
Type: | Integer |
Default: |
First-generation recognizers: 4 when the sample interval is <=500ms, but reducing to 1 as the sample interval is increased Second-, third-, and fourth-generation recognizers: 6 |
Required: | No |
Configuration Section: | TaskName |
Example: | FullAnalysisRate=4
|
See Also: | SampleInterval |