QFinger is a state-of-the-art
fingerprint image quality assessment software library .Net API, which is
available as a SDK for integration into third party applications and featured in
all fingerprint solutions, and wherever fingerprint quality assurance according
to international standards and recommendations is needed, whether captured live
or from paper, as in a back-record conversion scenario.
QFinger uses a set of complex and specific algorithms to assess the overall quality and usability of a fingerprint image in order to assure the highest possible quality of the captured data to be stored into a central register and to assure suitability for any AFIS system.
◊ Extensive QA Feedback ◊
QFinger implements a generic measure of fingerprint quality and is capable of determining:
• Crop information – the fingerprint bounding rectangle within the original image.
• Ridge and ridge flow QA.
• Reliable and unreliable image areas.
• Reliable and unreliable minutiae’s.
• Core and delta presence and position relative to the upper left corner of the cropped area.
• Minutiae count.
• Overall quality metric.
• NFIQ metric.
◊ World-class Additional Functionality ◊
• Live QA by analysing real-time buffer from fingerprint scanner.
• Adequate cropping of fingerprint.
• Slap Segmentation.
• 2 finger segmentation (optional).
• Sequence check.
• Authenticity check (1:10 fps to prevent any double captured fingerprint).
• FBI certified WSQ compression.
• Generating ANSI / NIST type records (2000, 2007 & 2008 XML).
• JPEG2000 support for 1000dpi images.
• RAW image file support - the raw pixel output of a fingerprint sensor.
◊ Reliable and Unreliable Images Areas
and Minutiae ◊
Areas based on the continuity of ridge flow across all regions of a finger image, and returns information based on the following factors:
• Proper size and resolution according to international standards.
• Core and delta presence.
• Global QA value.
• Image brightness, contrast and clarity to adequately determine ridges, minutiae, bifurcations, etc.
to ensure a fingerprint that is neither a too dark nor a too light.
• Smudged image due to movements, scratches / obstacles, wet fingers, etc.
◊ Quality Overlay ◊
All areas of interest are returned with certain assessment values which can be freely configured to be shown in the customer preferred colours to support the operator in a quality assured and guided live capture process or for batch processing with defined rules.
◊ Different Colours ◊
Can be assigned for:
• Too light image areas (bad area).
• Too dark image areas (bad area).
• Unreliable image areas (bad area).
• Good image areas (good area).
◊ Default Overlay Functionality
• Core and delta location.
• Reliable and unreliable minutiae.
• Good and bad areas of the image.
• Original image as a background (optional).
◊ Default Controls for Overlay
• Colour of core and delta.
• Colour of minutiae.
• Colour of reliable and unreliable minutiae.
• Colour of good and bad areas of the image.
• Colour of original image as a background (optional).
◊ Quality Metric ◊
QFinger calculates an overall score between 0 and 100, taking into account all available image assessment information and adequately determines the quality of the fingerprint.
◊ Quality Metric Thresholds ◊
Thresholds come with default values that are optimized according to empirical data vs. FBI compliant test sets. The user has the
ability to set his own thresholds according to his requirements.
• Bad (red colour).
• Acceptable (yellow colour).
• Good (green colour).
◊ QFinger - Advanced Fingerprint Pre and Post Processor
Designed to process low quality fingerprints, reconstruct ridge flows,
preserve features of interest and remove distortion and semi-permanent features
producing a significantly improved output that enhances matching performance.
A recent study evaluated the effect of the pre-processor on NFIQ quality scoring of fingerprints. Before application of the
pre-processor 90% of fingerprints scored at the lowest quality levels of 3-5, after pre-processing this reduced to just 10%.
By modelling the underlying ridge pattern even with low quality images, removing semi-permanent features (such as scratches and scars) and reducing linear and nonlinear distortions the
pre-processor module can improve matching performance of most fingerprint matching products.
Below you can see an improved fingerprint image by the pre- / post-processor.
This innovative pre-processor is another milestone to ensure high quality enrolments and the highest possible accuracy for any AFIS due to the use of quality assured fingerprints.
Below you can see the QFinger engine utilized in a customized fingerprint module optimized for guided workflows
and the quality metric determined and overlaid over the captured fingerprint.
This is directly stored as a FBI certified WSQ file. If the user clicks on one of the captured fingerprint images at the bottom, the Quality Overlay and additional information if shown. Otherwise the user gets instant feedback via the colour scheme if everything is ok or not, to
ensure the best usability and optimized enrolment time.