Photo Quality and Facial Recognition

qFaceQFace is an ICAO and ISO 19794-5 conform photo capture (for full frontal image),  quality assurance and post processing solution. It is either used standalone or is available for integrators as a .NET library or as entire API with full or partial GUI, as preferred by the customer. Additionally QFace is featured in all live and form based enrolment systems and any other solution that requires capturing a quality assured photo within the application workflow.

QFace evaluates live video frames in real-time as well as still images, and provides immediate quality feedback guiding the operator during the capture process to acquire the best possible final image in quality and compliance with the international standards. The live view inherits real time QA and live image overlay for the face finder, eye finder, nose finder and mouth finder and is constantly tracking all these parameters. Additionally the ICAO cropping frame is directly shown in the live view as an overlay. All overlays have direct QA feedback with regards to the displayed self explanatory colours (green and red). Auto capture functionality is supported as well, in this mode the snapshot of a person will be captured automatically triggering the camera capture function as soon as all quality conditions are met.

The image whether from live video or from file is then immediately cropped according to the eye position and ICAO specified eye distance in pixels and is automatically rotated according to the eye angle and automatic image enhancement. Adjustment of the image parameters and auto levels then takes place. Images can get stored directly in any specified format including JEPG2000 with inherited support for the NIST token image.

Additional features that make QFace unique in the market are features for dedicated post processing workflows, feathered edge image support as well as manual and automatic background removal of known and unknown backgrounds. Constant QA feedback for every quality check and process guide the operator through the workflow and assuring the quality criteria defined by ICAO or according to customer specific requirements are met.

Key Features

Architecture and Multiple Domain

• Standalone.
• Full ontegrated.
• .Net library (SDK).
• Complete module with open APIs for integration and data exchange.
• Highly flexible and customizable.

DIRFT Technology

DIRFT is an acronym for ‘Do It Right the First Time’, as we strongly believe that IT systems must inherently promote the guided, automatic, correct and quality assured capture of biometric information at the earliest possible point in time, as opposed to correcting information later in the enrolment process.

Guided Workflow and Powerful Business Logic

QFace can be operated automatically and semi-automatically to deliver the best of both worlds and to also cater effectively for exceptional cases. Constant QA checks in signal colours at every stage guide the operator to a quality assured result. Wherever possible the DIRFT approach is used i.e. the ICAO cropping frame shown in the live image is immediately turned to red and it is not possible to take an image if the cropping frame is outside of the image area.

Manual Post Processing Support

QFace allows for for manual intervention by the operator in the QA process and he can:

• Photo switch via thumbnails or automatic selection of photo with best QA (optional).
• Locate the eyes manually and perform ICAO compliant cropping based on selected eyes.
• Adjust image enhancement levels and settings manually.
• Manually post process non-uniform and unknown backgrounds for background removal.

Hardware Independence

A variety of cameras are supported out of the box by QFace:

• Digital cameras (Canon , Nikon, etc.)
• Industrial cameras.
• WebCams.
• Any DirectShow enabled camera for real time video processing, camera configuration and automatic / manual snapshot function.

Other Features

• Thumbnail preview.
• Load from file support.
• Full image type support including JPEG 2000.
• NIST / ICAO token face support.
• Overall QA score for every action performed.
• Freely configurable for specific business logics and processes.
• “Feathered Edge” image generation support.

ICAO / ISO 19794-5 Mandatory Requirements - Supported

Below is a list of featured checks. However the operator needs to provide a proper enrolment environment to cater for various occasions that can be checked but not corrected by the QA engine. i.e. equal illumination for no shadows on face and background or to advise the applicant to not wear tinted glasses etc.

Is Frontal: It can be divided into three checks for1.) IsFrontal_Pan - 2.) IsFrontal_Til- 3.) IsFrontal_Roll. Additionally rotation of head is less than +/-5 degrees with respect to Pan/Tilt axis & less than +/-8 degrees with regards to roll angle.

(ISO 7.2.2).

Only One Face Visible: Only one face has to be visible

(ISO 7.2.4)

Is Lighting Uniform: Lighting equally distributed in face

(ISO 7.2.7)

No Shadows: No strong shadows on the face

(ISO 7.2.8)

No Hotspots: No bright areas of light reflected from face

(ISO 7.2.10)

Good Exposure: There is no over or under exposure

(ISO 7.3.2)

Horizontally Centered Face: Face is cantered between 45% and 55% of the image width

(ISO 8.3.2)

GoodVerticalFacePosition: 0.5 image height < eyes line < 0.7 image height

(ISO 8.3.3)

WidthOfHead: Width of head is between 50% and 75% of image width

(ISO 8.3.4)

LengthOfHead: Length of head between 60% to 90% of image height

(ISO 8.3.5)

Resolution: At least 90 pixels from eye centre to eye centre

(ISO 8.4.1)

GoodGrayScaleProfile: Intensity resolution of at least 7 bits


ColorDepth: The minimum number of bits for colour image representations will be 8 bits per channel


ColorSpace: Input image colour space is YUV422, 8-bit greyscale or RGB24


EyesNotRed: Both eyes pupils are not detected as red

(ISO 7.3.4)

NoShadowsInAroundEyeSockets: No strong shadows in and around eye sockets

(ISO 7.2.9)

NoTintedGlasses: Person either wears no glasses or glasses are not tinted

(ISO 7.2.11)

IsSharp: Face area fits the focus and depth in field characteristics

(ISO 7.3.3)

HeadFullyVisible: Head is fully visible based on estimated chin/crown/ear1/ear2 positions

(ISO 8.3.1)

HasNaturalSkinColour: Face region has natural colours

(ISO 7.3.4 )

Best Practice Recommendations - Supported and Supported in future releases

Below is a list of featured optional checks out of the best practice recommendations.

IsFrontalBP: Rotation of head is less than +/-5 degrees from frontal in every direction (roll, pitch, yaw).

(ISO 7.2.2 and A 2.2)

ResolutionBP: 120 pixels distance from eye to eye.

(ISO A 3.1.1)

ImageWidthToHeightBP: Ratio (height/width) between 1.25 and 1.34.

(ISO A 3.2.1)

WidthOfHeadBP: Image width to face width ratio between 1.4 and 2.0.

(ISO A 3.2.2)

LengthOfHeadBP: Face length from 70% to 80% of the image height.

(ISO A 3.2.3)

NoShadowsOnBackgroundBP: No strong shadows on background.

(ISO A 2.4.2)

IsBackgroundUniformBP: Background is uniform.

(ISO A 2.4.3)

EyesGazeFrontalBP: Eyes are looking frontal to camera - Not currently supported - future release

(ISO 7.2.3 and A 2.2.1)

EyesOpenBP: Both eyes of the person are open - Not currently supported - future release

(ISO 7.2.3 and A 2.2.1)

MouthClosedBP: Mouth is closed - Not currently supported - future release

(ISO 7.2.3 and A 2.2.1)



QFaceIn the screenshot on the left, you can see a customized application / photo module utilizing the QFace engine. It features subsets of the QFace functionality, such as:

• Live QA feedback to support the operator in the process.
• Live finder and tracker of face, eye, nose with live QA (red and green)
• Live ICAO cropping frame with inherited business logic (red and green)
• Automatic cropping, rotating and photo enhancement.
• Auto capture with thumbnails and constant QA feedback.
• Summary of QA checks for easy readability and Global QA feedback.
• Camera control.
• Manual post processing options.

Copyright 2010 | Privacy Policy | Terms and Conditions | Disclaimer