Attributes

objects

angles

Head rotation angles:

• yaw: rotation around vertical Y-axis;
• pitch: rotation around horizontal Z-axis;
• roll: rotation around horizontal X-axis.

age

Age of a person in an image

bbox

The rectangle that represents face bounds in the image. The bounds are calculated relative to the coordinates of the original image.

class

Object class name, e.g., face or body

confidence

Numerical value of detection confidence

emotions

• emotion: 7 basic emotions: angry, disgusted, scared, happy, neutral, sad, surprised;
• confidence: the numerical value of manifestation of each estimated emotion. Value range: 0 to 1.

fitter

• fitter_type: type of set of anthropometric points. The fda set contains 21 points and provides high accuracy in a wide range of facial angles (up to the full profile), in contrast to the previous sets;
• keypoints: coordinates of anthropometric points;
• left_eye: left eye coordinates;
• right_eye: right eye coordinates.

gender

Gender of a person in an image

id

The ordinal number of a face in an image

liveness

• confidence: numerical value of confidence that the image belongs to a real person. Value range: 0 to 1;
• value: verdict: REAL - the face image belongs to a real person, FAKE - the face image doesn't belong to a real person.

mask

• value: verdict: true - masked person, false - unmasked person;
• confidence: the numerical value of confidence that a person in the image is/isn’t wearing a mask. Value range: 0 to 1.

quality

• quality:
  • qaa (Quality Assessment Algorithm) contains the following data:
    • totalScore: numerical value for overall image quality score in points from 0 to 100;
    • isSharp: boolean value for image sharpness;
    • sharpnessScore: numerical value for sharpness assessment in points from 0 to 100;
    • isEvenlyIlluminated: boolean value for image illumination uniformity;
    • illuminationScore: numerical value for score of illumination uniformity in points from 0 to 100;
    • noFlare: boolean value for presence/absence of image flares;
    • isLeftEyeOpened: boolean value for position of left eye (open/closed);
    • leftEyeOpennessScore: numerical value for eye openness score in points from 0 to 100;
    • isRightEyeOpened: boolean value for position of right eye (open/closed);
    • rightEyeOpennessScore: numerical value for eye openness score in points from 0 to 100;
    • isBackgroundUniform: boolean value for background uniformity;
    • backgroundUniformityScore: numerical value for background uniformity score in points from 0 to 100;
    • isDynamicRangeAcceptable: boolean value shows that dynamic range of image intensity in face area exceeds/does not exceed the value of 128;
    • dynamicRangeScore: numerical value for score of dynamic range of intensity in points from 0 to 100;
    • isEyesDistanceAcceptable: boolean value for acceptable/unacceptable distance between the eyes;
    • eyesDistance: numerical value of distance between eyes in px;
    • isNotNoisy: boolean value for presence/absence of image noise;
    • noiseScore: numerical value of image noise in points from 0 to 100;
    • isMarginsAcceptable: boolean value for acceptable/unacceptable margins;
    • marginInnerDeviation: numerical value of inner deviation in px;
    • marginOuterDeviation: numerical value of outer deviation in px;
    • isNeutralEmotion: boolean value for presence/absence of neutral emotions;
    • neutralEmotionScore: numerical value for score of neutral emotions in points from 0 to 100;
    • notMasked: boolean value for presence/absence of medical mask;
    • notMaskedScore: numerical value of confidence that a person in the image isn’t wearing a medical mask in points from 0 to 100;
    • hasWatermark: boolean value for presence/absence of watermark in the image;
    • watermarkScore: numerical value of confidence that the image contains a watermark in points from 0 to 100;
    • isRotationAcceptable: boolean value for acceptable/ unacceptable yaw, pitch and roll angles;
    • maxRotationDeviation: numerical value for max deviation of yaw, pitch and roll angles.

$template

Biometric template coded in base64

template_size

Size of a biometric template (vector length)

verification

distance

Distance between compared vectors of biometric face templates. The lower the value, the higher the degree of similarity.

fa_r (False Acceptance Rate)

False acceptance rate shows the system resistance to false acceptance errors. Such an error occurs when the biometric system recognizes a new face as previously detected one, i.e. images of different people are mistaken for images of the same person. This rate is measured by the number of false-acceptance recognitions divided by the total number of recognition attempts.

fr_r (False Rejection Rate)

False rejection occurs when a system fails to recognize a previously detected face, i.e. two images of the same person are mistaken for images of different people. The rate shows the percentage of recognition attempts with false rejection result.

score

The attribute shows the degree of similarity of faces from 0 (0%) to 1 (100%). A high degree of similarity means that two biometric templates belong to the same person.