The uniqueness of the human fingerprint is considered to be one of the most reliable characteristics for personal identification. Considering that we leave them on nearly all of the surfaces we touch, they become valuable to those in law enforcement for identifying perpetrators of a crime. However, the matching of a single fingerprint with the millions that have been cataloged proves to be a difficult task. This study presents an alternate method to fingerprint recognition by way of a spatial re-sampling of the pattern through concentric circles. With this approach, the concentric circular samples have rotation invariant features while a translation is dependent only on the location of the circles' center. The resulting circles are then correlated with those from the known set to obtain a collection of the most probable matches. This technique has shown exceptional results when comparing various binary test patterns as well as synthetic binary fingerprint images but is unable to recognize unenhanced greyscale fingerprint images.
Chang, David, "Fingerprint recognition through circular sampling" (1999). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus