This publication describes a fingerprint-matching algorithm that uses a fusion of minutiae and pattern-correlation matching. Instead of extracting minutiae points, the algorithm extracts small patches from the fingerprint image and transforms them into rotationally-invariant vectors. The algorithm divides each image to be evaluated, herein called an “enrolled” image, into “M” number of patches with a sliding distance of one (1) pixel. The algorithm also extracts “N” number of random patches from a stored image, herein called a “verify” image, and calculates the similarity between the “verify” rotationally-invariant vectors and the “enrolled” rotationally-invariant vectors. At this stage, the algorithm merges vectors from the “enrolled” images using a rotation and translation matrix and drops redundant vectors based on a quality score. The outcome of matching is the number or matching blocks that show similar translation and rotation. Finally, the algorithm generates a “Yes-or-No” outcome based on a predetermined threshold number of matching blocks.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.