A Modified Thinning Algorithm For Minutiae Feature Extraction Of Fingerprint Images On FPGA
Critical and indispensable step in biometric fingerprint image processing is thinning. The speed and reliability of
thinning process are more important for the fingerprint. In this paper a fast fingerprint thinning algorithm is introduced. We
have modified the thinning algorithm which was proposed by Saleh, Eldin and Wadah .The unique feature of our
modified algorithm that directly works on gray scale images after normalization process not the binarized one where
binarization of fingerprint causes many spurious minutiae and removes some important features and algorithm makes use of
parallel processing of each individual image segments of size 8×8 pixels to speed up the execution without segmentation and
ridge extraction. This paper presents the implementation of algorithm on Spartan- 3E XC3S100E FPGA as an appropriate
solution for portable devices and for low-cost consumer market. Experimental results shows thinning of fingerprint image
are obtained in 321ms when an image of 256×256 pixel is analyzed.
Keywords- Biometrics, Fingerprint Identification, Image Processing.