Segmentation and enhancement of fingerprint images based on automatic threshold calculations

Abstract

A new approach to fingerprint image segmentation and feature extraction is proposed to improve the implementation of Automated Fingerprint Identification System (AFIS) based on automatic threshold values. The process starts by partitioning the fingerprint image manually into 16 × 16 pixels blocks. For each block, a local threshold is calculated using its mean, variance and coherence. Then, statistical analysis is performed to find the optimal threshold value for each block. This threshold is then used to extract a foreground of the fingerprint image from the background. Later, the foreground is enhanced using a newly developed technique called filling-in-the-gap process to fill in the gaps in the foreground and eliminate any unwanted handwritten annotations in the image. The current method was evaluated using the NIST-14 database and showed reliable results on different quality images.