A Novel Authentication Approach Based on Level 2 Minutiae-based Feature Extraction Using Gabor Filter

The fingerprint is considered to be a robust and stable authentication approach and the minutiae present in the fingerprints provide the most reliable information about a person. But still, the quality of enrolled fingerprint images makes the task more challenging (Bharadwaj et al. in EURASIP J Image Video Process, 1–28, 2014, [1]). Due to the improper acquisition of samples (angle of impression and pressure), the fingerprints suffer poor quality, noises, interclass, and intraclass variability of input samples (Olsen et al. in IET Biom 5:47–64, 2016, [2]) Hence both pre and post-processing are necessary for reliable extraction of feature descriptors. In these scenarios, The Minutiae-based approach is considered to be robust for feature extraction and verification. In this paper, we are using the fingerprints from the FVC2002 public fingerprint Database for studying and analyzing our approach for feature extraction and robust minutiae detection in the fingerprint. The results of our novel approach have provided successfully detected minutiae on processed and enhanced skeletonized fingerprint images using Iterative fingerprint enhancement with the Gabor filter approach. The generated minutiae score from this approach can be used for minutiae authentication.