/OpenAFQA

Automated Fingermark Quality Assessment - Tools, methods, resources and more.

Primary LanguagePythonMIT LicenseMIT

OpenAFQA

The repository for developing Automated Fingermark Quality Assessment methods.

This library is a work in progress. We will update it as we continue our development on the topic.

Contents

  • AFQA Toolbox. The continuously developed Automated Fingermark Quality Assessment toolbox consists of a collection of commonly used algorithms for friction ridge preprocessing and feature extraction. Also included are Python implementations or wrappers of fingerprint/fingermark quality assessment methods.
  • Toolbox examples. Practical examples, where the usage of the toolbox is demonstrated.
  • Experiments. Code for various publications, related to automated fingermark quality assessment.

Toolbox Installation

  1. Set up a Python environment by installing packages in requirements.txt
  2. To use the minutiae extraction wrappers, download and compile the code. For more information on this, see minutiae extraction README.
  3. The pre-trained models for our quality assessment methods can be downloaded from this link. Either copy the downloaded models into afqa_toolbox/resources/ or use a custom path at initialization.
  4. The toolbox can be installed locally by running python setup.py install or python setup.py develop if you want to modify its contents.

Latest updates

  • (October, 2022) The AFQA Ensemble models - This release includes two approaches to AFQA: (a) a classic predictive pipeline with preprocessing, feature extraction and feature vector creation steps, as well as (b) a deep learning model which processes raw fingermark images. We also include a fusion method that combines multiple quality scores into one. Read more about this contribution in our Knowledge-Based Systems publication.
  • (September, 2021) AFQA Toolbox - We released the AFQA toolbox, a collection of algorithms for processing fingerprint/fingermark images, written in Python. We also provided the initial baseline models for quality assessment of fingermarks. Read more about this contribution in our BIOSIG publication.

References

If you use our open-source software, please consider citing:

T. Oblak, R. Haraksim, P. Peer, L. Beslay. 
Fingermark quality assessment framework with classic and deep learning ensemble models. 
Knowledge-Based Systems, Volume 250, 2022    

T. Oblak, R. Haraksim, L. Beslay, P. Peer. 
Fingermark Quality Assessment: An Open-Source Toolbox. 
In proceedings of the International Conference of the Biometrics Special Interest Group (BIOSIG), pp. 159-170, 2021.