Biometric fingerprint quality detection
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
NFIQ2 Initial Commit Aug 10, 2017
OpenCV
biomdi
complianceTestSet
libFRFXLL
DISCLAIMER.md
LICENSE.md
Makefile
README.md
version.mak

README.md

NFIQ 2.0

NFIQ 2.0 is a revision of the open source NIST Finger Image Quality (NFIQ). In 2004, NIST developed the first publicly available fingerprint quality assessment tool NFIQ. Advances in fingerprint technology since 2004, necessitated an update to NFIQ. As such, development of NFIQ 2.0 was initiated in 2011 as collaboration between National Institute of Standards and Technology (NIST) and Federal Office for Information Security (BSI) and Federal Criminal Police Office (BKA) in Germany as well as research and development entities, MITRE, Fraunhofer IGD, Hochschule Darmstadt (HAD) and Secunet.

NFIQ 2.0 provides a higher resolution quality score, in range of 0-100 according to the international biometric sample quality standard ISO/IEC 29794-1:2016 (as opposed to 1-5), lower computation complexity, as well as support for quality assessment in mobile platform.

The major innovation of NFIQ was linking image quality to operational recognition performance. This had several immediate benefits; it allowed quality values to be tightly defined and then numerically calibrated. This, in turn, allowed for the standardization needed to support a worldwide deployment of fingerprint sensors with universally interpretable image qualities. NFIQ 2.0 is the basis for a revision of the Technical Report ISO/IEC 29794-4 Biometric sample quality -- Part 4:Finger image data:2010 into an international standard. Specifically, NFIQ quality features are being formally standardized as part of ISO/IEC 29794-4 Biometric sample quality -- Part 4: Finger image data and NFIQ source code serves as the reference implementation of the standard.

Operationally, NFIQ has increased the reliability, accuracy, and interoperability of fingerprint recognition systems by identifying the samples that are likely to cause recognition failure.

If you would like more information please read the NFIQ 2.0 Report.

How to Build

If all requirements have been met, building is as simple as:

make
sudo make install

Library Path for libbiomdi will need to be set to run the binary.

Requirements

  • A supported operating system:

    • RHEL/CentOS >= 6.x
    • Ubuntu
    • macOS >= 10.11
  • System packages

System Packages

Some modules require system packages that may not be installed by default on all operating systems. Package names are listed below for RHEL/CentOS, Ubuntu and macOS (via MacPorts). Other operating systems may use similarly-named packages.

Name RHEL/CentOS MacPorts Ubuntu
gcc gcc n/a (requires Command Line Tools) gcc-6
g++ gcc-c++ n/a (requires Command Line Tools) g++-6
CMAKE cmake cmake cmake
OpenCV opencv-devel Build from source included libopencv-dev

*** A minimum version of OpenCV 2.4.2 is required and OpenCV 3.0 => is not supported at this time. ***

*** MacOS users running 10.12 must use OpenCV 2.4.13.2 ***

####OpenCV build steps if building locally

 mkdir libOpenCV && cd libOpenCV && cmake -D CMAKE_MAKE_PROGRAM=make ../OpenCV && make opencv_core opencv_ts opencv_imgproc opencv_highgui opencv_flann opencv_features2d opencv_calib3d opencv_ml opencv_video opencv_objdetect opencv_contrib opencv_nonfree opencv_gpu opencv_photo opencv_stitching opencv_videostab
 sudo make install

If you do not have root access or do not wish to install OpenCV you will need to set the Library Path for OpenCV

Communication

If you found a bug and can provide steps to reliably reproduce it, or if you have a feature request, please open an issue. Other questions may be addressed to the project maintainers.

License

NFIQ is released in the public domain. See the LICENSE for details.