Skip to content

Neural networks used in LUNA ID#

In LUNA ID, neural networks provide efficient and accurate processing of faces in images and video streams. Neural networks are stored in .plan files.

The table below shows the .plan files used in LUNA ID for Android and iOS and functionality that the files cover.

.plan file

Feature name

Description

More information

angle_estimation_flwr_arm.plan

Head pose estimation

Determines person head rotation angles in 3D space, that is pitch, yaw, and roll.

Best shot quality estimation

ags_angle_estimation_flwr_arm.plan

Best shot quality estimation

Evaluates image quality to choose the best image before descriptor extraction. The BestShotQuality estimator consists of two components - AGS (Approximate Garbage Store) and Head Pose.

ags_estimation_flwr_arm.plan

AGS estimation

Determines the source image score for further descriptor extraction and matching.

cnn52m_arm.plan

Descriptor generation from an image

Stores a compact set of packed properties as well as some helper parameters used to extract these properties from the source image.

Descriptor

cnn52m_cpu.plan (in LUNA ID for Android only)

cnn59m_arm.plan

cnn59m_cpu.plan (in LUNA ID for Android only)

eye_status_estimation_flwr_arm.plan

Eye state

Determines the eye state: open, closed, occluded.

Eye estimation

eyes_estimation_flwr8_arm.plan

Eye state estimation

Determines the following eye state and keypoints:

  • Eye state: open, closed, occluded.
  • Precise eye iris location as an array of landmarks.
  • Precise eyelid location as an array of landmarks.

FaceDet_v2_first_arm.plan

Face detection

Detects a face in an image and shows a rectangular area around the detected face.

The neural networks should be launched consequently.

Face detection

FaceDet_v2_second_arm.plan

FaceDet_v2_third_arm.plan

glasses_estimation_flwr_arm.plan

Image glasses estimation

Determines whether a person is currently wearing glasses.

Glasses estimation

model_subjective_quality_v1_arm.plan

Image quality estimation

Determines an image quality by the following criteria:

  • The image is blurred.
  • The image is underexposed, that is, too dark.
  • The image is overexposed, that is, too light.
  • The face in the image is illuminated unevenly and there is a great difference between dark and light regions.
  • The image contains flares on face, that is, too specular.

Image quality estimation

model_subjective_quality_v2_arm.plan

Configuration options of the supported features are stored in the faceengine.conf file. The file locates in "data/faceengine.conf" in current working directory.

Warning! We do not recommend that you change any configuration settings from default ones as these settings affect performance and output results of your application.

For more information about the settings stored in the faceengine.conf file, see: