F. Benhammadi, H. Hentous, K. Bey Beghdad, and M. Aissani (Algeria)
Biometrics, pattern recognition, fingerprints matching.
This paper discusses a new technique to perform fingerprint ridge-based matching approach avoiding pre alignment features extracted from the two fingerprints images. An orientation-dependent tessellation features attached to each minutia is defined to facilitate this. Thus, our approach uses the minutia information to generate the cells of the square grid and the weighted features using weighting factor to guarantee reliable matching results. These features are invariant to translation and rotation what permits to avoid the fingerprints pair alignment stage. The algorithm matching was tested on the fingerprint database DB1 used in FVC2002 and the results are promising.
Important Links:
Go Back
