OVERVIEW
About this Guide
This guide aims to offer a straightforward introduction into setting up the face recognition feature.
RESOURCES
If you need further information on to how to handle the settings.json file you can refer to the server KB articles or contact support.
REQUIREMENTS:
Profile Pictures:
Quantity: At least 5 profile pictures of the subject. The eyes and mouth are the most important details.
Specifications:
Single portraits only. No group pictures. No cropped pictures.
Black and white pictures are not supported.
.HEIC format is not supported. Use .JPEG preferably.
Resolution should be at least 720 x 1080 pixels
Pictures can’t have a microphone or object in front of the face
Pictures can’t have any covering over the eyes like sunglasses
Hats should not shade the face
Media Content:
Duration: At least 5 seconds of clear apparition of the subject in the media content.
CONFIGURATION:
Step 1 - Make a directory and place the pictures in it
Make a directory named after the subject with the referred pictures in a shared Media folder.
Important Reminder: When performing the search the query must be identical to the name of the directory. This includes case letters and special characters.
Example:
Search:
Directory:
Step 2 - Configure the settings path
Within the settings.json file add the path to the directory that contains the subject's directories:
"FACES_PATHS": {
"user2024": [
"/share/media/faces/"
]
},
The final path to the subject directory should look like this:
/share/media/faces/<subjects-name>
You can add as many users as required.