Device Set-Up Requirements

Constraints

There are several constraints set in order to obtain an optimal performance:

  • The camera is placed in a hallway with a minimum length of 3 meters.

  • Users are not allowed to wear face attributes other than eyeglasses or a mask.

  • If there is a recognition failure when the user is wearing glasses or a mask, the user is expected to take off the glasses or mask in order to provide more features for recognition.

  • Users’ face needs to be evenly illuminated.

  • Users need to walk at an average walking pace.

Camera Set-Up

  • Image resolution, there is no minimum requirement for image resolution, as long as the face resolution is 125 x 125 pixels. Higher-resolution images require a larger minimum face size. For example, if you have 360p camera resolution, the distance camera to people should be less than 45 cm.

  • Place the camera in good lighting not too dark or bright and with no backlight.

  • Camera height, between 1.5 m - 1.75 m.

  • Frame rate, the recommended frame rate is 30 fps. (at least 25 fps)

  • Encoder Bitrate, the recommended encoder bitrate is 3 Mbps.

  • Codec, the codec should be h.264.

  • Give a marker place approximately less than 3 m from the camera to take pictures.

Marker Example

Trigger

Some ideas that can be used to trigger photo submissions to the API:

  • Printed QR Code

  • RFID

  • NFC

  • Ultrasonic Module

  • Infrared Module

  • Face location and minimal size

Last updated

Was this helpful?