This document summarizes a fingerprint post processing algorithm that performs minutia extraction and false minutia removal. It first describes preprocessing steps including binarization, orientation estimation, and thinning. It then outlines the minutia detection process which uses a rotation mask method and crossing number algorithm to identify bifurcations and terminations. Finally, it details a false minutia removal process that compares local patches around detected minutia to templates to identify and remove minutia located on image edges or in close proximity, which are likely false detections.
8. False Minutia Removal
• Create an image of minutia
• Take a chunk around a minutia point
• Compare that image with given mask image
• If these are at the edge then remove that
minutia
12. False Minutia Removal Cont’d
• Take an array, size is nearly equal to distance
between two lines. We take25x25 because
distance between lines is around 14 pixel.
• Find the minutia and take a chunk of 25x25.
• If there is minutia within this window then
remove both minutia.
• If there is bifurcation then remove the
termination and bifurcation because it is a
spike.
14. False Minutia Removal Cont’d
• Like we take window around the termination
we also take the window around the
bifurcation
• Find if there is more than one bifurcation
within the window remove both bifurcation
because it is hole.
19. References
• Rohit Singh , Utkarsh Shah , Vinay Gupta
Indian Institute of technology, Kanpur
• Fingerprint Image Enhancement and Minutiae
Extraction, Raymond Thai, The University of
Western Australia