FinetuneFaceDatabase
Performs a final, optional, stage in the training process that optimizes the neural network to improve recognition accuracy.
This action acts on a single face database, which must meet several requirements:
- The database must have been created with fine-tuning enabled (see the
finetuning
parameter in the action CreateFaceDatabase). - The database must contain a reasonable number of faces. OpenText recommends more than 50.
- The database must contain multiple training images for most or all of the faces.
NOTE: OpenText recommends that you fine-tune the neural network only if your training data is similar to the media that you will ingest. For example, if you are planning to ingest video captured from a CCTV camera, the training data should ideally have been captured from CCTV. In some cases, using action=FinetuneFaceDatabase
can result in overfitting and reduce recognition accuracy. If you find that accuracy is reduced, you can revert to the default neural network by setting DisableFineTuning=TRUE
in your face recognition task.
OpenText recommends running the action BuildAllFaces before running FinetuneFaceDatabase
, because untrained images are not used to optimize the neural network.
NOTE: The optimized network is used automatically, but you must set the Database parameter when you configure the face recognition task in your session configuration. Otherwise, Media Server uses a generic face recognition network that is suitable for all databases, rather than the optimized network produced by this action.
Running this action does not require a GPU.
Type: asynchronous
Parameter | Description | Required |
---|---|---|
database
|
The name of a database to process. | Yes |
Example
/action=FinetuneFaceDatabase&database=politicians
Response
This action is asynchronous, so Media Server always returns success accompanied by a token. You can use this token with the QueueInfo action to retrieve the status of your request.
The FinetuneFaceDatabase
action can take a significant amount of time to complete (over an hour with a large database). It writes progress information to the application log.