For proper face recognition and comparison the images uploaded to the reference face database (see Selecting a way to upload an image to the reference face database) are to meet the following requirements:

  1. There must be a full face image and the resolution is to be HD (1280×720 pixels) or FullHD (1920×1080 pixels). 
  2. The image is to be sharp, not blurred, steady illuminated and not exposed.
  3. The distance between eyes is not to be less than 60 pixels (at least 30 pixels for the Huawei recognition module).
  4. A person in the image is to occupy the greater part of the frame - so as the person's head and shoulders are to be visible (it is not recommended to use waist height images).

For VisionLabs recognition module the size of the photo added to the database is configured in the CascadeDetector::SizeHint settings section of the config.xml file located in the <Intellect installation directory>\Modules\FaceRecognition\VisionLabs\data.