Reconstructing Fingerprints from Minutiae Points
Most fingerprint-based biometric systems store the minutiae template of a user in the database. It has been traditionally assumed that the minutiae template of a user does not reveal any information about the original fingerprint. In this paper, we challenge this notion and show that three levels of information about the parent fingerprint can be elicited from the minutiae template alone, viz., 1) the orientation field information, 2) the class or type information, and 3) the friction ridge structure. The orientation estimation algorithm determines the direction of local ridges using the evidence of minutiae triplets. The estimated orientation field, along with the given minutiae distribution, is then used to predict the class of the fingerprint. Finally, the ridge structure of the parent fingerprint is
generated using streamlines that are based on the estimated orientation field. Line Integral Convolution is used to impart texture to the ensuing ridges, resulting in a ridge map resembling the parent fingerprint. The salient feature of this noniterative method to generate ridges is its ability to preserve the minutiae at specified locations in the reconstructed ridge map. Experiments using a commercial fingerprint matcher suggest that the reconstructed ridge structure bears close resemblance to the parent fingerprint.
A fingerprint is an oriented texture pattern consisting of ridges and valleys present on the tip of an individualâ„¢s
finger. The ridges exhibit various types of imperfections, called minutiae (minor details in fingerprints). Among a total of 150 different minutiae types, the ridge ending and ridge bifurcation are the most stable points in a fingerprint. The distribution of these points in a fingerprint has been observed to be unique across individuals. In fact, this distribution is claimed to be unique to each finger of an individual. Thus, most automatic fingerprint authentication systems do not store the raw fingerprint image of a user in its entirety during enrollment. Rather, a template consisting of a set of salient features (e.g., singular points, such as core and delta, and ridge anomalies, such as minutiae) from the fingerprint image is stored in the database. Since the template, by definition, is a compact description of the biometric sample, it is not expected to reveal significant information about the original data. Therefore, template generation algorithms have been traditionally assumed to be one-way algorithms. However, Hill designed a technique to determine the fingerprint structure from the minutiae template alone. He assumed that a fingerprint template stores the coordinates of the core and delta points (if present) along with the minutiae points. His technique utilized the location of the singular points to derive the orientation map of the fingerprint based on the method proposed in. This orientation map was then used by a heuristic line drawing algorithm to generate a sequence of splines passing through the minutiae points. Hill demonstrated his reconstruction scheme on a small database of 25 arch type fingerprints. His scheme also predicted the shape of the fingerprint (i.e., its class) using a neural network classifier consisting of 23 input neurons, 13 hidden neurons, and four output neurons (corresponding to the four major fingerprint classes, viz., A, L, R, and W). However, the classification performance was observed to be rather low (an error rate of 28.9 percent on a small set of 242 fingerprints). Similarly, in the face domain, Adler demonstrated that a face image could be reconstructed from face templates using a hill climbing attack.