Q1 : What is your standard of ideal photo for enrolment?
A1: Please see  Xnap Face Recognition Best Practice document.

Q2: What is the image size of the camera supported by Xnap Box ? XGA, UXGA?
A2: Maximum resolution supported is 1920 x 1080 pixel, which FHD format.
Camera with small resolution will also be supported automatically, say XGA (1024×768)

Q3: Is there a specific port the rasp should be streamed on?
A3: Xnap Box is not picky about RTSP port to use, as long as that port is not occupied by other services. Default port of RTSP is 554

Q4: In order to see if faces are being captured, what mode do I use on the Xnap Box to grab face image? HTTP Push or Presentation?
A4: Please use “HTTP Push”.
Presentation mode is designed to test if the connection between the camera and HTTP caller is viable without face/object detection.
This option must be saved and restart the Xnap Box to make Presentation mode effective. From the Main Menu of Home Page, “Face Stream” / “Object Stream” menu is renamed as “Presentation Mode”.
It is useful for troubleshooting when no face/object was detected from HTTP caller point of view. It could be camera viewpoint problem (be specific, too wide angle, captured faces are smaller than minimum face size), image quality issues (out of focus, too blurry or noisy images), or simply network connection issue between the camera and HTTP caller.
When this radio button was selected, live video of the current selected source URL will be compressed to 320×240 pixel and export through HTTP protocol without passing through the face/object detection engine.
In case that no image stream was received by the HTTP caller, then Administrator should check the network connection and the camera video output.