The main aim of this finished project was to apply global models to fingerprint images for robust extraction of local features. These so-called minutiae features are used within classic pattern recognition algorithms for fingerprint matching (recognition, authentication). The direction field of a fingerprint is a crucial parameter for extracting minutiae features. Nevertheless many fingerprint images are of such poor quality, that the direction of the field can not be extracted for certain regions in the image. On the other hand it has been shown that if one can properly "guess" the direction, it is possible to apply enhancement algorithms which adaptively improve the clarity of ridges and furrows of such regions. In order to do this "guess work" computationally, a model for the directional field of a fingerprint was applied during the extraction process. In another concept the extracted parameters of the directional field model were employed for fingerprint classification.
Surinder Ram received a M.Sc. in Telematics in 2006 from Graz University of Technology. He was a research assistant at the Institute for Computer Graphics and Vision at Graz University of Technology. He finished his PhD-Thesis focused on Fingerprint-Ridge Orientation Modelling in 2010.
Related master student works:
Robert Hödl worked on a segmentation algorithm for slap fingerprint segmentation
Martin Hamker worked on a fingerprint image quality estimation method
This paper presents a novel approach in segmenting multiple fingerprints from an image. A combination of two-staged mean shift and ellipse-fitting algorithms as well as an elaborate subsequent set of rules is used to segment the single fingertip images. Extensive experimental evaluations demonstrate the success of the approach.
This paper proposes a statistical model for fingerprint ridge orientations. The active fingerprint ridge orientation model (AFROM) iteratively deforms to fit the orientation field (OF) of a fingerprint. The main application of the method is the OF estimation in noisy fingerprints as well as the interpolation and extrapolation of larger OF parts. Fingerprint OFs are represented by Legendre Polynomials. We evaluated both,the generalisation as well as the prediction capability of the proposed method. These evaluations assess our method very good results.
Ram S, Bischof H, Birchbauer J: Presented at 3rd IAPR/IEEE International Conference on Biometrics (ICB), Alghero, Italy, Jun. 2-5 2009. PDF
Curvature Preserving Fingerprint Ridge Orientation Smooting using Legendre Polynomials
Smoothing fingerprint ridge orientation involves a principal discrepancy. Too little smoothing can result in noisy orientation fields (OF), too much smoothing will harm high curvature areas, especially singular points (SP). In this paper we present a fingerprint ridge orientation model based on Legendre polynomials.
Ram S, Bischof H, Birchbauer J: Presented at IEEE Computer Society Workshop on Biometrics, associated with IEEE CVPR 2008, Anchorage, USA. Received Best Student Paper Award.PDF,Video
Detection of Singularities in Fingerprint Images using Linear Phase Portraits
In this paper we present a model based approach for the detection of singular points. The presented method exploits the geometric nature of linear differential equation systems.
Ram S, Bischof H, Birchbauer J: Published as a chapter in Handbook of Remote Biometrics, Springer Series Advances in Pattern Recognition, pp. 349-362. PDF,DOI
A robust model based algorithm for detection of singularities in fingerprint images
The performance of fingerprint recognition is heavily depending on the reliable extraction of singularities. Common algorithms are based on a Poincare-Index estimation which is a numerical method. These algorithms ignore the topology of the underlying data and are only robust when certain heuristics and rules are applied. In this paper we present a model based approach for the detection of singular points. The presented method exploits the geometric nature of linear differential equation systems.
Ram S, Bischof H, Birchbauer J: Presented at 12th Computer Vision Winter Workshop (CVWW) 2007, St. Lambrecht, Austria, Feb 6-8. PDF