Fingerprint orientation field regularisation via multi-target regression

• Published in: Electronics letters Vol. 52; no. 13; pp. 1118 - 1120
• Main Authors: Lin, Lu, Liu, Eryun, Wang, Lianghao, Zhang, Ming
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 23.06.2016
• Subjects: Algorithms; Computer simulation; deep neural networks; Electronics; feature extraction; fingerprint feature extraction; fingerprint feature recognition; fingerprint identification; fingerprint orientation field regularisation; Fingerprints; FVC2002 database; FVC2004 database; FVC2006 database; Image and vision processing and display technology; Mathematical models; multitarget regression model; neural nets; Neural networks; Orientation; orientation field estimation; orientation patches; Regression; regression analysis; fingerprint identification; FVC2006 database; regression analysis; fingerprint orientation field regularisation; FVC2004 database; multitarget regression model; fingerprint feature extraction; orientation field estimation; FVC2002 database; feature extraction; fingerprint feature recognition; orientation patches; neural nets; deep neural networks
• ISSN: 0013-5194; 1350-911X; 1350-911X
• DOI: 10.1049/el.2015.4483

Orientation field estimation is a key step in fingerprint feature extraction and recognition. A complete orientation field estimation algorithm usually consists of two steps, i.e. initial orientation field estimation and post regularisation. In this Letter, a multi-target regression model to regularise the initial orientation field is proposed. A large number of orientation patches with simulated noises, together with their regression targets are fed to a deep neural networks to train a multi-target regression model. For a given initial orientation field at testing stage, a refined orientation field is obtained by applying the regression model in patch-wise and then combining all predicted patches. Experimental results on FVC2002, FVC2004 and FVC2006 databases show remarkable performance compared with state of the art algorithms. Our algorithm is also highly efficient and easy to implement.