Izadi, M. HamedDrygajlo, Andrzej2013-10-032013-10-032013-10-032013https://infoscience.epfl.ch/handle/20.500.14299/96090Poor quality of fingerprint data is one of the major problems concerning latent fingerprint matching in forensic applications. Local quality of fingerprint plays a very important role in this application field to ensure high recognition performance. Al- though big progress has been made in matching of fingerprints using local minutiae descriptors, in particular Minutia Cylinder- Code (MCC), automatic latent fingerprint matching continues to be a challenge. Previously we proposed a matching algo- rithm that uses minutiae information encoded by MCC with in- tegrated local quality measures associated to each MCC called cylinder quality measures. In our previous work, cylinder qual- ity measures for latent case have been proposed by combining the subjective qualities of individual minutiae involved. In this paper, we propose an alternative method to estimate the cylin- der quality measures directly from fingerprint quality maps, in particular ridge clarity maps, by taking into account the num- ber of involving minutiae as well. Integration of MCC with the proposed cylinder quality measures was evaluated through ex- periments on the latent fingerprint database NIST SD27, which show clear improvements in the identification performance of latent fingerprints of ugly quality.text::conference output::conference proceedings::conference paper