Intrinsic Limitations of Fingerprint Orientation Estimation

Estimation of orientation field is a crucial issue when processing fingerprint samples. Many subsequent fingerprint processing steps depend on reliable and accurate estimations. Algorithms for such estimations are usually evaluated against ground truth data. As true ground truth is usually not available, human experts need to mark-up ground truth manually. However, the accuracy and the reliability of such mark-ups for orientation fields have not been investigated yet. Mark-ups produced by six humans allowed insights into both aspects. A Root Mean Squared Error of about 7° against true ground truth can be achieved. Reproducibility between two mark-ups of a single dactyloscopic expert is at the same precision. We concluded that the accuracy of human experts is competitive to the best algorithms evaluated at FVC-ongoing.

[1]  R. A. Hicklin,et al.  Changes in latent fingerprint examiners' markup between analysis and comparison. , 2015, Forensic science international.

[2]  Christoph Busch,et al.  Deep expectation for estimation of fingerprint orientation fields , 2017, 2017 IEEE International Joint Conference on Biometrics (IJCB).

[3]  R. A. Hicklin,et al.  Accuracy and reliability of forensic latent fingerprint decisions , 2011, Proceedings of the National Academy of Sciences.

[4]  Qijun Zhao,et al.  Model Based Separation of Overlapping Latent Fingerprints , 2012, IEEE Transactions on Information Forensics and Security.

[5]  Dario Maio,et al.  Improving Fingerprint Orientation Extraction , 2011, IEEE Transactions on Information Forensics and Security.

[6]  D. Maio,et al.  Semi-automatic enhancement of very low quality fingerprints , 2009, 2009 Proceedings of 6th International Symposium on Image and Signal Processing and Analysis.

[7]  R. A. Hicklin,et al.  Repeatability and Reproducibility of Decisions by Latent Fingerprint Examiners , 2012, PloS one.

[8]  Christoph Busch,et al.  Predicting Dactyloscopic Examiner Fingerprint Image Quality Assessments , 2015, 2015 International Conference of the Biometrics Special Interest Group (BIOSIG).

[9]  R. A. Hicklin,et al.  Measuring What Latent Fingerprint Examiners Consider Sufficient Information for Individualization Determinations , 2014, PloS one.

[10]  Raffaele Cappelli,et al.  SFinGe : an Approach to Synthetic Fingerprint Generation , 2004 .

[11]  Elham Tabassi,et al.  Semantic Conformance Testing Methodology for Finger Minutiae Data , 2009 .

[12]  Davide Maltoni,et al.  Benchmarking Local Orientation Extraction in Fingerprint Recognition , 2010, 2010 20th International Conference on Pattern Recognition.

[13]  Carsten Gottschlich,et al.  Performance evaluation of fingerprint orientation field reconstruction methods , 2015, 3rd International Workshop on Biometrics and Forensics (IWBF 2015).

[14]  R. A. Hicklin,et al.  Interexaminer variation of minutia markup on latent fingerprints. , 2016, Forensic science international.

[15]  Christoph Busch,et al.  Semantic Conformance Testing Method for Finger Minutiae Data , 2009, BIOSIG.

[16]  Axel Munk,et al.  Robust Orientation Field Estimation and Extrapolation Using Semilocal Line Sensors , 2009, IEEE Transactions on Information Forensics and Security.

[17]  Anil K. Jain,et al.  Orientation Field Estimation for Latent Fingerprint Enhancement , 2013, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[18]  Melissa K. Taylor,et al.  Markup Instructions for Extended Friction Ridge Features , 2013 .