A Fault-Tolerant Approach for Detection of Singular Points in Noisy Fingerprint Images
Singular point detection is one of the most crucial processes in fingerprint recognition systems. Singular points are used for fingerprint classification, matching and alignment. This paper presents a method for reliable detection of singular points, which is largely insensitive to the degradation of fingerprint quality. The approach involves two phases, wherein, first phase detects the singular points, which operates on the quadrant change information. The second phase involves the analysis and extraction of the locations having high probability of the existence of singular points. The second phase uses orientation reliability measure of the filtered fingerprint image. The spurious singular points are detected, and thereby eliminated. This model is tested on selected noisy images from a publicly available (Cross Match Verifier 300 sensor) fingerprint database scanned at 500 dpi. The experimental results show that the approach effectively eliminates the spurious singular points in the noisy images.
Keywords: Singular Points, Spurious, Quadrant Change, Reliability, Minimum Inertia, Maximum Inertia, Euclidean Distance
Download Full-Text
