Generation of Synthetic Alterations

STRANGE (SynThetic fingeRprint AlteratioNs GEnerator) is a novel approach for the generation of synthetic alterations on fingerprint images. Three of the most frequent alterations, namely obliteration by burning, distortion by central rotation and by Z-cut are modeled by designing quite realistic alteration shapes and by including scars and transitions between native and altered regions.

Why a generator of synthetic fingerprint alterations?

An important issue in the study of automatic detection/recognition algorithm of altered fingerprints is the lack of public databases containing altered fingerprints and the corresponding pre-altered version available to researchers to perform systematic tests. This lack often stymies researchers to work in this area. In order to treat with the altered fingerprints problems it becomes necessary to have databases that include the more frequent alterations and a pre-altered impression of the fingerprint itself.
STRANGE software would be an useful tool for generating, at zero cost, databases of altered fingerprints, thus to be used in the development and optimization of new algorithms for detecting or matching altered fingerprints. In particular altered databases can be useful for:

• tuning the parameters when developing algorithms for the detection or matching of alterations;
• performing a training set needed in the learning step of new recognition approach for matching altered fingerprints;
• performing ad hoc algorithm for classifying the alterations: actually the input parameters of the generator tool allow to explicitly control the different kind of alterations to be generated. Databases for each alteration can be easily generated and used during tests and optimizations;
• performing ad hoc algorithm for detecting the level of alterations.

Version History

• Version 1.0 - first version released
• Version 1.1 - some minor bugs have been corrected