Robust orientation field estimation in fingerprint images with broken ridge lines

A good and reliable orientation field (OF) estimation is of great importance in fingerprint image processing. This paper presents an OF estimation approach that is based on traced ridge and furrow lines. The method considers pieces of flow information from a larger region and provides robustness against disturbances caused, e.g. by temporary scars, creases, contamination, wetness or dryness of the finger. The performance of the method is evaluated in simulation runs with artificial ridge interruptions and compared to the gradients based method and a multiscale directional operator.