A combined statistical and structural approach for fingerprint image postprocessing

A postprocessing approach that uses both statistical and structural information to eliminate the false minutiae in skeleton fingerprint images is described. The attributes of fingerprint ridges are presented by statistics, and these attributes are used to describe the structure of the false minutiae. The proposed approach analyzes the entire minutia structures, not just in a small window, much like the way humans analyze a fingerprint pattern. To verify the effectiveness of the approach, ten processed images with different shapes were contrasted with rolled fingerprint cards. False minutiae made up 36% of the total minutiae before processing and only about 4% after processing. Moreover, the ridge separation and continuity are good throughout the fingerprint image.<>