The first step in face recognition is face detection, therefore all of the parameters used to tune face detection also apply to face recognition. You can use four additional parameters to tune the performance of face recognition.
Description | Default value | Example | |
---|---|---|---|
Database
|
The database that Image Server compares each face against. | Match against all face databases |
Image Server compares faces in an image against the models in the |
MaxRecognitionResults
|
The maximum number of potential matches returned for each detected face. | 5 |
For each face, Image Server returns the two matches with the highest confidence ratings. |
RecognitionThreshold
|
The lowest confidence rating that a match can have and be returned by Image Server. Below this value, Image Server does not report the match. | No lower limit |
Image Server returns matches with confidence ratings higher than 0.6 (up to the number of results specified by |
Identifier
|
A comma-separated list of identifiers, each of which identifies a specific face in the database. Image Server restricts comparisons to the specified faces only. To use this parameter you must set the |
Match against all faces in all databases |
Image Server compares faces to only the database models with the identifiers |
|